style fix

* auto intergrate joyimage model

* joyimage pipeline

* train

* ready

* styling

* joyai-image docs

* update readme

* pr review

.gitignore

@@ -2,6 +2,7 @@
 /models
 /scripts
 /diffusers
+/.vscode
 *.pkl
 *.safetensors
 *.pth

README.md

@@ -7,6 +7,7 @@
 [![open issues](https://isitmaintained.com/badge/open/modelscope/DiffSynth-Studio.svg)](https://github.com/modelscope/DiffSynth-Studio/issues)
 [![GitHub pull-requests](https://img.shields.io/github/issues-pr/modelscope/DiffSynth-Studio.svg)](https://GitHub.com/modelscope/DiffSynth-Studio/pull/)
 [![GitHub latest commit](https://badgen.net/github/last-commit/modelscope/DiffSynth-Studio)](https://GitHub.com/modelscope/DiffSynth-Studio/commit/)
+[![Discord](https://badgen.net//discord/members/Mm9suEeUDc)](https://discord.gg/Mm9suEeUDc)
 
 [切换到中文版](./README_zh.md)
 
@@ -31,8 +32,13 @@ We believe that a well-developed open-source code framework can lower the thresh
 
 > DiffSynth-Studio has undergone major version updates, and some old features are no longer maintained. If you need to use old features, please switch to the [last historical version](https://github.com/modelscope/DiffSynth-Studio/tree/afd101f3452c9ecae0c87b79adfa2e22d65ffdc3) before the major version update.
 
-> Currently, the development personnel of this project are limited, with most of the work handled by [Artiprocher](https://github.com/Artiprocher). Therefore, the progress of new feature development will be relatively slow, and the speed of responding to and resolving issues is limited. We apologize for this and ask developers to understand.
-- **January 19, 2026**: Added support for [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) and [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) models, including training and inference capabilities. [Documentation](/docs/en/Model_Details/Wan.md) and [example code](/examples/mova/) are now available.
+> Currently, the development personnel of this project are limited, with most of the work handled by [Artiprocher](https://github.com/Artiprocher) and [mi804](https://github.com/mi804). Therefore, the progress of new feature development will be relatively slow, and the speed of responding to and resolving issues is limited. We apologize for this and ask developers to understand.
+
+- **April 24, 2026** We add support for Stable Diffusion v1.5 and SDXL, including inference, low VRAM inference, and training capabilities. For details, please refer to the [documentation](/docs/en/Model_Details/Stable-Diffusion.md), [documentation](/docs/en/Model_Details/Stable-Diffusion-XL.md) and [example code](/examples/stable_diffusion/).
+
+- **April 14, 2026** JoyAI-Image open-sourced, welcome a new member to the image editing model family! Support includes instruction-guided image editing, low VRAM inference, and training capabilities. For details, please refer to the [documentation](/docs/en/Model_Details/JoyAI-Image.md) and [example code](/examples/joyai_image/).
+
+- **March 19, 2026**: Added support for [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) and [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) models, including training and inference capabilities. [Documentation](/docs/en/Model_Details/Wan.md) and [example code](/examples/mova/) are now available.
 
 - **March 12, 2026**: We have added support for the [LTX-2.3](https://modelscope.cn/models/Lightricks/LTX-2.3) audio-video generation model. The features includes text-to-audio/video, image-to-audio/video, IC-LoRA control, audio-to-video, and audio-video inpainting. We have supported the complete inference and training functionalities. For details, please refer to the [documentation](/docs/en/Model_Details/LTX-2.md) and [code](/examples/ltx2/).
 
@@ -40,6 +46,9 @@ We believe that a well-developed open-source code framework can lower the thresh
 
 - **March 2, 2026** Added support for [Anima](https://modelscope.cn/models/circlestone-labs/Anima). For details, please refer to the [documentation](docs/en/Model_Details/Anima.md). This is an interesting anime-style image generation model. We look forward to its future updates.
 
+<details>
+<summary>More</summary>
+
 - **February 26, 2026** Added full and lora training support for the LTX-2 audio-video generation model. See the [documentation](/docs/en/Model_Details/LTX-2.md) for details.
 
 - **February 10, 2026** Added inference support for the LTX-2 audio-video generation model. See the [documentation](/docs/en/Model_Details/LTX-2.md) for details. Support for model training will be implemented in the future.
@@ -67,9 +76,6 @@ We believe that a well-developed open-source code framework can lower the thresh
     - [Differential LoRA Training](/docs/zh/Training/Differential_LoRA.md): This is a training technique we used in [ArtAug](https://www.modelscope.cn/models/DiffSynth-Studio/ArtAug-lora-FLUX.1dev-v1), now available for LoRA training of any model.
     - [FP8 Training](/docs/zh/Training/FP8_Precision.md): FP8 can be applied to any non-training model during training, i.e., models with gradients turned off or gradients that only affect LoRA weights.
 
-<details>
-<summary>More</summary>
-
 - **November 4, 2025** Supported the [ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B) model, which is trained based on Wan 2.1 and supports generating corresponding actions based on reference videos.
 
 - **October 30, 2025** Supported the [meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video) model, which supports text-to-video, image-to-video, and video continuation. This model uses the Wan framework for inference and training in this project.
@@ -295,6 +301,129 @@ Example code for Z-Image is available at: [/examples/z_image/](/examples/z_image
 
 </details>
 
+#### Stable Diffusion: [/docs/en/Model_Details/Stable-Diffusion.md](/docs/en/Model_Details/Stable-Diffusion.md)
+
+<details>
+
+<summary>Quick Start</summary>
+
+Running the following code will quickly load the [AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5) model for inference. VRAM management is enabled, the framework automatically controls parameter loading based on available VRAM, requiring a minimum of 2GB VRAM.
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion import StableDiffusionPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars, high quality, detailed",
+    negative_prompt="blurry, low quality, deformed",
+    cfg_scale=7.5,
+    height=512,
+    width=512,
+    seed=42,
+    rand_device="cuda",
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+</details>
+
+<details>
+
+<summary>Examples</summary>
+
+Example code for Stable Diffusion is available at: [/examples/stable_diffusion/](/examples/stable_diffusion/)
+
+|Model ID|Inference|Low VRAM Inference|Full Training|Full Training Validation|LoRA Training|LoRA Training Validation|
+|-|-|-|-|-|-|-|
+|[AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5)|[code](/examples/stable_diffusion/model_inference/stable-diffusion-v1-5.py)|[code](/examples/stable_diffusion/model_inference_low_vram/stable-diffusion-v1-5.py)|[code](/examples/stable_diffusion/model_training/full/stable-diffusion-v1-5.sh)|[code](/examples/stable_diffusion/model_training/validate_full/stable-diffusion-v1-5.py)|[code](/examples/stable_diffusion/model_training/lora/stable-diffusion-v1-5.sh)|[code](/examples/stable_diffusion/model_training/validate_lora/stable-diffusion-v1-5.py)|
+
+</details>
+
+#### Stable Diffusion XL: [/docs/en/Model_Details/Stable-Diffusion-XL.md](/docs/en/Model_Details/Stable-Diffusion-XL.md)
+
+<details>
+
+<summary>Quick Start</summary>
+
+Running the following code will quickly load the [stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0) model for inference. VRAM management is enabled, the framework automatically controls parameter loading based on available VRAM, requiring a minimum of 6GB VRAM.
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion_xl import StableDiffusionXLPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder_2/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer/"),
+    tokenizer_2_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer_2/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars",
+    negative_prompt="",
+    cfg_scale=5.0,
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+</details>
+
+<details>
+
+<summary>Examples</summary>
+
+Example code for Stable Diffusion XL is available at: [/examples/stable_diffusion_xl/](/examples/stable_diffusion_xl/)
+
+|Model ID|Inference|Low VRAM Inference|Full Training|Full Training Validation|LoRA Training|LoRA Training Validation|
+|-|-|-|-|-|-|-|
+|[stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0)|[code](/examples/stable_diffusion_xl/model_inference/stable-diffusion-xl-base-1.0.py)|[code](/examples/stable_diffusion_xl/model_inference_low_vram/stable-diffusion-xl-base-1.0.py)|[code](/examples/stable_diffusion_xl/model_training/full/stable-diffusion-xl-base-1.0.sh)|[code](/examples/stable_diffusion_xl/model_training/validate_full/stable-diffusion-xl-base-1.0.py)|[code](/examples/stable_diffusion_xl/model_training/lora/stable-diffusion-xl-base-1.0.sh)|[code](/examples/stable_diffusion_xl/model_training/validate_lora/stable-diffusion-xl-base-1.0.py)|
+
+</details>
+
 #### FLUX.2: [/docs/en/Model_Details/FLUX2.md](/docs/en/Model_Details/FLUX2.md)
 
 <details>
@@ -596,6 +725,143 @@ Example code for FLUX.1 is available at: [/examples/flux/](/examples/flux/)
 
 </details>
 
+#### ERNIE-Image: [/docs/en/Model_Details/ERNIE-Image.md](/docs/en/Model_Details/ERNIE-Image.md)
+
+<details>
+
+<summary>Quick Start</summary>
+
+Running the following code will quickly load the [PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image) model and perform inference. VRAM management is enabled, and the framework will automatically control the loading of model parameters based on available VRAM. The model can run with a minimum of 3GB VRAM.
+
+```python
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("output.jpg")
+```
+
+</details>
+
+<details>
+
+<summary>Examples</summary>
+
+Example code for ERNIE-Image is available at: [/examples/ernie_image/](/examples/ernie_image/)
+
+| Model ID | Inference | Low VRAM Inference | Full Training | Full Training Validation | LoRA Training | LoRA Training Validation |
+|-|-|-|-|-|-|-|
+|[PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image)|[code](/examples/ernie_image/model_inference/ERNIE-Image.py)|[code](/examples/ernie_image/model_inference_low_vram/ERNIE-Image.py)|[code](/examples/ernie_image/model_training/full/ERNIE-Image.sh)|[code](/examples/ernie_image/model_training/validate_full/ERNIE-Image.py)|[code](/examples/ernie_image/model_training/lora/ERNIE-Image.sh)|[code](/examples/ernie_image/model_training/validate_lora/ERNIE-Image.py)|
+|[PaddlePaddle/ERNIE-Image-Turbo](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo)|[code](/examples/ernie_image/model_inference/ERNIE-Image-Turbo.py)|[code](/examples/ernie_image/model_inference_low_vram/ERNIE-Image-Turbo.py)|—|—|—|—|
+
+</details>
+
+#### JoyAI-Image: [/docs/en/Model_Details/JoyAI-Image.md](/docs/en/Model_Details/JoyAI-Image.md)
+
+<details>
+
+<summary>Quick Start</summary>
+
+Running the following code will quickly load the [jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit) model and perform inference. VRAM management is enabled, and the framework will automatically control the loading of model parameters based on available VRAM. The model can run with a minimum of 4GB VRAM.
+
+```python
+from diffsynth.pipelines.joyai_image import JoyAIImagePipeline, ModelConfig
+import torch
+from PIL import Image
+from modelscope import dataset_snapshot_download
+
+# Download dataset
+dataset_snapshot_download(
+    dataset_id="DiffSynth-Studio/diffsynth_example_dataset",
+    local_dir="data/diffsynth_example_dataset",
+    allow_file_pattern="joyai_image/JoyAI-Image-Edit/*"
+)
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+
+pipe = JoyAIImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="transformer/transformer.pth", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/model*.safetensors", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="vae/Wan2.1_VAE.pth", **vram_config),
+    ],
+    processor_config=ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+# Use first sample from dataset
+dataset_base_path = "data/diffsynth_example_dataset/joyai_image/JoyAI-Image-Edit"
+prompt = "将裙子改为粉色"
+edit_image = Image.open(f"{dataset_base_path}/edit/image1.jpg").convert("RGB")
+
+output = pipe(
+    prompt=prompt,
+    edit_image=edit_image,
+    height=1024,
+    width=1024,
+    seed=0,
+    num_inference_steps=30,
+    cfg_scale=5.0,
+)
+
+output.save("output_joyai_edit_low_vram.png")
+```
+
+</details>
+
+<details>
+
+<summary>Examples</summary>
+
+Example code for JoyAI-Image is available at: [/examples/joyai_image/](/examples/joyai_image/)
+
+| Model ID | Inference | Low VRAM Inference | Full Training | Full Training Validation | LoRA Training | LoRA Training Validation |
+|-|-|-|-|-|-|-|
+|[jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit)|[code](/examples/joyai_image/model_inference/JoyAI-Image-Edit.py)|[code](/examples/joyai_image/model_inference_low_vram/JoyAI-Image-Edit.py)|[code](/examples/joyai_image/model_training/full/JoyAI-Image-Edit.sh)|[code](/examples/joyai_image/model_training/validate_full/JoyAI-Image-Edit.py)|[code](/examples/joyai_image/model_training/lora/JoyAI-Image-Edit.sh)|[code](/examples/joyai_image/model_training/validate_lora/JoyAI-Image-Edit.py)|
+
+</details>
+
 ### Video Synthesis
 
 https://github.com/user-attachments/assets/1d66ae74-3b02-40a9-acc3-ea95fc039314
@@ -835,41 +1101,43 @@ graph LR;
 
 Example code for Wan is available at: [/examples/wanvideo/](/examples/wanvideo/)
 
-| Model ID | Extra Args | Inference | Full Training | Full Training Validation | LoRA Training | LoRA Training Validation |
-|-|-|-|-|-|-|-|
-|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
-|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
-|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
-|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
-|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
-|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
-|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
-|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
-|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
-|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
-|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
-|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
-|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
-|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
-|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](/examples/wanvideo/model_inference/LongCat-Video.py)|[code](/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
-|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
-|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
-|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
-|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
-|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
-|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
-|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
-|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
-|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
-|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
-| [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) | `input_image` | [code](/examples/mova/model_inference/MOVA-360p-I2AV.py) | [code](/examples/mova/model_training/full/MOVA-360P-I2AV.sh) | [code](/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py) | [code](/examples/mova/model_training/lora/MOVA-360P-I2AV.sh) | [code](/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py) |
-| [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) | `input_image` | [code](/examples/mova/model_inference/MOVA-720p-I2AV.py) | [code](/examples/mova/model_training/full/MOVA-720P-I2AV.sh) | [code](/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py) | [code](/examples/mova/model_training/lora/MOVA-720P-I2AV.sh) | [code](/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py) |
+| Model ID | Extra Inputs | Inference | Low VRAM Inference | Full Training | Validation After Full Training | LoRA Training | Validation After LoRA Training |
+|-|-|-|-|-|-|-|-|
+|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
+|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
+|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
+|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
+|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
+|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
+|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
+|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
+|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
+|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
+|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
+|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
+|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
+|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
+|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
+|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
+|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
+|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
+|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
+|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
+|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
+|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
+|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
+|[openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py)|
+|[openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py)|
+|[Wan-AI/WanToDance-14B (global model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-global.py)|
+|[Wan-AI/WanToDance-14B (local model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-local.py)|
 
 </details>
 
@@ -1027,3 +1295,9 @@ https://github.com/Artiprocher/DiffSynth-Studio/assets/35051019/b54c05c5-d747-47
 https://github.com/Artiprocher/DiffSynth-Studio/assets/35051019/59fb2f7b-8de0-4481-b79f-0c3a7361a1ea
 
 </details>
+
+## Contact Us
+
+|Discord：https://discord.gg/Mm9suEeUDc|
+|-|
+|<img width="160" height="160" alt="Image" src="https://github.com/user-attachments/assets/29bdc97b-e35d-4fea-88d6-32e35182e458" />|

README_zh.md

@@ -7,6 +7,7 @@
 [![open issues](https://isitmaintained.com/badge/open/modelscope/DiffSynth-Studio.svg)](https://github.com/modelscope/DiffSynth-Studio/issues)
 [![GitHub pull-requests](https://img.shields.io/github/issues-pr/modelscope/DiffSynth-Studio.svg)](https://GitHub.com/modelscope/DiffSynth-Studio/pull/)
 [![GitHub latest commit](https://badgen.net/github/last-commit/modelscope/DiffSynth-Studio)](https://GitHub.com/modelscope/DiffSynth-Studio/commit/)
+[![Discord](https://badgen.net//discord/members/Mm9suEeUDc)](https://discord.gg/Mm9suEeUDc)
 
 [Switch to English](./README.md)
 
@@ -31,8 +32,13 @@ DiffSynth 目前包括两个开源项目：
 
 > DiffSynth-Studio 经历了大版本更新，部分旧功能已停止维护，如需使用旧版功能，请切换到大版本更新前的[最后一个历史版本](https://github.com/modelscope/DiffSynth-Studio/tree/afd101f3452c9ecae0c87b79adfa2e22d65ffdc3)。
 
-> 目前本项目的开发人员有限，大部分工作由 [Artiprocher](https://github.com/Artiprocher) 负责，因此新功能的开发进展会比较缓慢，issue 的回复和解决速度有限，我们对此感到非常抱歉，请各位开发者理解。
-- **2026年1月19日** 新增对 [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) 和 [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) 模型的支持，包括完整的训练和推理功能。[文档](/docs/zh/Model_Details/Wan.md)和[示例代码](/examples/mova/)现已可用。
+> 目前本项目的开发人员有限，大部分工作由 [Artiprocher](https://github.com/Artiprocher) 和 [mi804](https://github.com/mi804) 负责，因此新功能的开发进展会比较缓慢，issue 的回复和解决速度有限，我们对此感到非常抱歉，请各位开发者理解。
+
+- **2026年4月24日** 我们新增对 Stable Diffusion v1.5 和 SDXL 的支持，包括推理、低显存推理和训练能力。详情请参考[文档](/docs/zh/Model_Details/Stable-Diffusion.md)和[示例代码](/examples/stable_diffusion/)。
+
+- **2026年4月14日** JoyAI-Image 开源，欢迎加入图像编辑模型家族！支持指令引导的图像编辑推理、低显存推理和训练能力。详情请参考[文档](/docs/zh/Model_Details/JoyAI-Image.md)和[示例代码](/examples/joyai_image/)。
+
+- **2026年3月19日** 新增对 [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) 和 [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) 模型的支持，包括完整的训练和推理功能。[文档](/docs/zh/Model_Details/Wan.md)和[示例代码](/examples/mova/)现已可用。
 
 - **2026年3月12日** 我们新增了 [LTX-2.3](https://modelscope.cn/models/Lightricks/LTX-2.3) 音视频生成模型的支持，模型支持的功能包括文生音视频、图生音视频、IC-LoRA控制、音频生视频、音视频局部Inpainting，框架支持完整的推理和训练功能。详细信息请参考 [文档](/docs/zh/Model_Details/LTX-2.md) 和 [示例代码](/examples/ltx2/)。
 
@@ -40,6 +46,9 @@ DiffSynth 目前包括两个开源项目：
 
 - **2026年3月2日** 新增对[Anima](https://modelscope.cn/models/circlestone-labs/Anima)的支持，详见[文档](docs/zh/Model_Details/Anima.md)。这是一个有趣的动漫风格图像生成模型，我们期待其后续的模型更新。
 
+<details>
+<summary>更多</summary>
+
 - **2026年2月26日** 新增对[LTX-2](https://www.modelscope.cn/models/Lightricks/LTX-2)音视频生成模型全量微调与LoRA训练支持，详见[文档](docs/zh/Model_Details/LTX-2.md)。
 
 - **2026年2月10日** 新增对[LTX-2](https://www.modelscope.cn/models/Lightricks/LTX-2)音视频生成模型的推理支持，详见[文档](docs/zh/Model_Details/LTX-2.md)，后续将推进模型训练的支持。
@@ -67,9 +76,6 @@ DiffSynth 目前包括两个开源项目：
     - [差分 LoRA 训练](/docs/zh/Training/Differential_LoRA.md)：这是我们曾在 [ArtAug](https://www.modelscope.cn/models/DiffSynth-Studio/ArtAug-lora-FLUX.1dev-v1) 中使用的训练技术，目前已可用于任意模型的 LoRA 训练。
     - [FP8 训练](/docs/zh/Training/FP8_Precision.md)：FP8 在训练中支持应用到任意非训练模型，即梯度关闭或者梯度仅影响 LoRA 权重的模型。
 
-<details>
-<summary>更多</summary>
-
 - **2025年11月4日** 支持了 [ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B) 模型，该模型基于 Wan 2.1 训练，支持根据参考视频生成相应的动作。
 
 - **2025年10月30日** 支持了 [meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video) 模型，该模型支持文生视频、图生视频、视频续写。这个模型在本项目中沿用 Wan 的框架进行推理和训练。
@@ -295,6 +301,129 @@ Z-Image 的示例代码位于：[/examples/z_image/](/examples/z_image/)
 
 </details>
 
+#### Stable Diffusion：[/docs/zh/Model_Details/Stable-Diffusion.md](/docs/zh/Model_Details/Stable-Diffusion.md)
+
+<details>
+
+<summary>快速开始</summary>
+
+运行以下代码可以快速加载 [AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5) 模型并进行推理。显存管理已启用，框架会自动根据剩余显存控制模型参数的加载，最低 2GB 显存即可运行。
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion import StableDiffusionPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars, high quality, detailed",
+    negative_prompt="blurry, low quality, deformed",
+    cfg_scale=7.5,
+    height=512,
+    width=512,
+    seed=42,
+    rand_device="cuda",
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+</details>
+
+<details>
+
+<summary>示例代码</summary>
+
+Stable Diffusion 的示例代码位于：[/examples/stable_diffusion/](/examples/stable_diffusion/)
+
+| 模型 ID | 推理 | 低显存推理 | 全量训练 | 全量训练后验证 | LoRA 训练 | LoRA 训练后验证 |
+|-|-|-|-|-|-|-|
+|[AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5)|[code](/examples/stable_diffusion/model_inference/stable-diffusion-v1-5.py)|[code](/examples/stable_diffusion/model_inference_low_vram/stable-diffusion-v1-5.py)|[code](/examples/stable_diffusion/model_training/full/stable-diffusion-v1-5.sh)|[code](/examples/stable_diffusion/model_training/validate_full/stable-diffusion-v1-5.py)|[code](/examples/stable_diffusion/model_training/lora/stable-diffusion-v1-5.sh)|[code](/examples/stable_diffusion/model_training/validate_lora/stable-diffusion-v1-5.py)|
+
+</details>
+
+#### Stable Diffusion XL：[/docs/zh/Model_Details/Stable-Diffusion-XL.md](/docs/zh/Model_Details/Stable-Diffusion-XL.md)
+
+<details>
+
+<summary>快速开始</summary>
+
+运行以下代码可以快速加载 [stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0) 模型并进行推理。显存管理已启用，框架会自动根据剩余显存控制模型参数的加载，最低 6GB 显存即可运行。
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion_xl import StableDiffusionXLPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder_2/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer/"),
+    tokenizer_2_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer_2/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars",
+    negative_prompt="",
+    cfg_scale=5.0,
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+</details>
+
+<details>
+
+<summary>示例代码</summary>
+
+Stable Diffusion XL 的示例代码位于：[/examples/stable_diffusion_xl/](/examples/stable_diffusion_xl/)
+
+| 模型 ID | 推理 | 低显存推理 | 全量训练 | 全量训练后验证 | LoRA 训练 | LoRA 训练后验证 |
+|-|-|-|-|-|-|-|
+|[stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0)|[code](/examples/stable_diffusion_xl/model_inference/stable-diffusion-xl-base-1.0.py)|[code](/examples/stable_diffusion_xl/model_inference_low_vram/stable-diffusion-xl-base-1.0.py)|[code](/examples/stable_diffusion_xl/model_training/full/stable-diffusion-xl-base-1.0.sh)|[code](/examples/stable_diffusion_xl/model_training/validate_full/stable-diffusion-xl-base-1.0.py)|[code](/examples/stable_diffusion_xl/model_training/lora/stable-diffusion-xl-base-1.0.sh)|[code](/examples/stable_diffusion_xl/model_training/validate_lora/stable-diffusion-xl-base-1.0.py)|
+
+</details>
+
 #### FLUX.2: [/docs/zh/Model_Details/FLUX2.md](/docs/zh/Model_Details/FLUX2.md)
 
 <details>
@@ -596,6 +725,143 @@ FLUX.1 的示例代码位于：[/examples/flux/](/examples/flux/)
 
 </details>
 
+#### ERNIE-Image: [/docs/zh/Model_Details/ERNIE-Image.md](/docs/zh/Model_Details/ERNIE-Image.md)
+
+<details>
+
+<summary>快速开始</summary>
+
+运行以下代码可以快速加载 [PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 3G 显存即可运行。
+
+```python
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("output.jpg")
+```
+
+</details>
+
+<details>
+
+<summary>示例代码</summary>
+
+ERNIE-Image 的示例代码位于：[/examples/ernie_image/](/examples/ernie_image/)
+
+| 模型 ID | 推理 | 低显存推理 | 全量训练 | 全量训练后验证 | LoRA 训练 | LoRA 训练后验证 |
+|-|-|-|-|-|-|-|
+|[PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image)|[code](/examples/ernie_image/model_inference/ERNIE-Image.py)|[code](/examples/ernie_image/model_inference_low_vram/ERNIE-Image.py)|[code](/examples/ernie_image/model_training/full/ERNIE-Image.sh)|[code](/examples/ernie_image/model_training/validate_full/ERNIE-Image.py)|[code](/examples/ernie_image/model_training/lora/ERNIE-Image.sh)|[code](/examples/ernie_image/model_training/validate_lora/ERNIE-Image.py)|
+|[PaddlePaddle/ERNIE-Image-Turbo](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo)|[code](/examples/ernie_image/model_inference/ERNIE-Image-Turbo.py)|[code](/examples/ernie_image/model_inference_low_vram/ERNIE-Image-Turbo.py)|—|—|—|—|
+
+</details>
+
+#### JoyAI-Image: [/docs/zh/Model_Details/JoyAI-Image.md](/docs/zh/Model_Details/JoyAI-Image.md)
+
+<details>
+
+<summary>快速开始</summary>
+
+运行以下代码可以快速加载 [jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 4G 显存即可运行。
+
+```python
+from diffsynth.pipelines.joyai_image import JoyAIImagePipeline, ModelConfig
+import torch
+from PIL import Image
+from modelscope import dataset_snapshot_download
+
+# Download dataset
+dataset_snapshot_download(
+    dataset_id="DiffSynth-Studio/diffsynth_example_dataset",
+    local_dir="data/diffsynth_example_dataset",
+    allow_file_pattern="joyai_image/JoyAI-Image-Edit/*"
+)
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+
+pipe = JoyAIImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="transformer/transformer.pth", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/model*.safetensors", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="vae/Wan2.1_VAE.pth", **vram_config),
+    ],
+    processor_config=ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+# Use first sample from dataset
+dataset_base_path = "data/diffsynth_example_dataset/joyai_image/JoyAI-Image-Edit"
+prompt = "将裙子改为粉色"
+edit_image = Image.open(f"{dataset_base_path}/edit/image1.jpg").convert("RGB")
+
+output = pipe(
+    prompt=prompt,
+    edit_image=edit_image,
+    height=1024,
+    width=1024,
+    seed=0,
+    num_inference_steps=30,
+    cfg_scale=5.0,
+)
+
+output.save("output_joyai_edit_low_vram.png")
+```
+
+</details>
+
+<details>
+
+<summary>示例代码</summary>
+
+JoyAI-Image 的示例代码位于：[/examples/joyai_image/](/examples/joyai_image/)
+
+|模型 ID|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|
+|[jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit)|[code](/examples/joyai_image/model_inference/JoyAI-Image-Edit.py)|[code](/examples/joyai_image/model_inference_low_vram/JoyAI-Image-Edit.py)|[code](/examples/joyai_image/model_training/full/JoyAI-Image-Edit.sh)|[code](/examples/joyai_image/model_training/validate_full/JoyAI-Image-Edit.py)|[code](/examples/joyai_image/model_training/lora/JoyAI-Image-Edit.sh)|[code](/examples/joyai_image/model_training/validate_lora/JoyAI-Image-Edit.py)|
+
+</details>
+
 ### 视频生成模型
 
 https://github.com/user-attachments/assets/1d66ae74-3b02-40a9-acc3-ea95fc039314
@@ -835,41 +1101,43 @@ graph LR;
 
 Wan 的示例代码位于：[/examples/wanvideo/](/examples/wanvideo/)
 
-|模型 ID|额外参数|推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
-|-|-|-|-|-|-|-|
-|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
-|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
-|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
-|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
-|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
-|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
-|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
-|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
-|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
-|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
-|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
-|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
-|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
-|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
-|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](/examples/wanvideo/model_inference/LongCat-Video.py)|[code](/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
-|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
-|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
-|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
-|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
-|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
-|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
-|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
-|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
-|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
-|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
-| [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) | `input_image` | [code](/examples/mova/model_inference/MOVA-360p-I2AV.py) | [code](/examples/mova/model_training/full/MOVA-360P-I2AV.sh) | [code](/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py) | [code](/examples/mova/model_training/lora/MOVA-360P-I2AV.sh) | [code](/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py) |
-| [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) | `input_image` | [code](/examples/mova/model_inference/MOVA-720p-I2AV.py) | [code](/examples/mova/model_training/full/MOVA-720P-I2AV.sh) | [code](/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py) | [code](/examples/mova/model_training/lora/MOVA-720P-I2AV.sh) | [code](/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py) |
+|模型 ID|额外参数|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|-|
+|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
+|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
+|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
+|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
+|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
+|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
+|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
+|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
+|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
+|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
+|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
+|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
+|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
+|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
+|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
+|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
+|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
+|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
+|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
+|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
+|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
+|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
+|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
+|[openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py)|
+|[openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py)|
+|[Wan-AI/WanToDance-14B (global model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-global.py)|
+|[Wan-AI/WanToDance-14B (local model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-local.py)|
 
 </details>
 
@@ -1029,3 +1297,9 @@ https://github.com/Artiprocher/DiffSynth-Studio/assets/35051019/b54c05c5-d747-47
 https://github.com/Artiprocher/DiffSynth-Studio/assets/35051019/59fb2f7b-8de0-4481-b79f-0c3a7361a1ea
 
 </details>
+
+## 联系我们
+
+|Discord：https://discord.gg/Mm9suEeUDc|
+|-|
+|<img width="160" height="160" alt="Image" src="https://github.com/user-attachments/assets/29bdc97b-e35d-4fea-88d6-32e35182e458" />|

diffsynth/configs/model_configs.py

@@ -307,6 +307,13 @@ wan_series = [
         "model_class": "diffsynth.models.wav2vec.WanS2VAudioEncoder",
         "state_dict_converter": "diffsynth.utils.state_dict_converters.wans2v_audio_encoder.WanS2VAudioEncoderStateDictConverter",
     },
+    {
+        # Example: ModelConfig(model_id="Wan-AI/WanToDance-14B", origin_file_pattern="global_model.safetensors")
+        "model_hash": "eb18873fc0ba77b541eb7b62dbcd2059",
+        "model_name": "wan_video_dit",
+        "model_class": "diffsynth.models.wan_video_dit.WanModel",
+        "extra_kwargs": {'has_image_input': True, 'patch_size': [1, 2, 2], 'in_dim': 36, 'dim': 5120, 'ffn_dim': 13824, 'freq_dim': 256, 'text_dim': 4096, 'out_dim': 16, 'num_heads': 40, 'num_layers': 40, 'eps': 1e-06, 'wantodance_enable_music_inject': True, 'wantodance_music_inject_layers': [0, 4, 8, 12, 16, 20, 24, 27], 'wantodance_enable_refimage': True, 'has_ref_conv': True, 'wantodance_enable_refface': False, 'wantodance_enable_global': True, 'wantodance_enable_dynamicfps': True, 'wantodance_enable_unimodel': True}
+    },
 ]
 
 flux_series = [
@@ -534,6 +541,22 @@ flux2_series = [
     },
 ]
 
+ernie_image_series = [
+    {
+        # Example: ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors")
+        "model_hash": "584c13713849f1af4e03d5f1858b8b7b",
+        "model_name": "ernie_image_dit",
+        "model_class": "diffsynth.models.ernie_image_dit.ErnieImageDiT",
+    },
+    {
+        # Example: ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors")
+        "model_hash": "404ed9f40796a38dd34c1620f1920207",
+        "model_name": "ernie_image_text_encoder",
+        "model_class": "diffsynth.models.ernie_image_text_encoder.ErnieImageTextEncoder",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.ernie_image_text_encoder.ErnieImageTextEncoderStateDictConverter",
+    },
+]
+
 z_image_series = [
     {
         # Example: ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors")
@@ -597,6 +620,13 @@ z_image_series = [
         "extra_kwargs": {"model_size": "0.6B"},
         "state_dict_converter": "diffsynth.utils.state_dict_converters.z_image_text_encoder.ZImageTextEncoderStateDictConverter",
     },
+    {
+        # To ensure compatibility with the `model.diffusion_model` prefix introduced by other frameworks.
+        "model_hash": "8cf241a0d32f93d5de368502a086852f",
+        "model_name": "z_image_dit",
+        "model_class": "diffsynth.models.z_image_dit.ZImageDiT",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.z_image_dit.ZImageDiTStateDictConverter",
+    },
 ]
 """
 Offical model repo: https://www.modelscope.cn/models/Lightricks/LTX-2
@@ -870,4 +900,75 @@ mova_series = [
         "model_class": "diffsynth.models.mova_dual_tower_bridge.DualTowerConditionalBridge",
     },
 ]
-MODEL_CONFIGS = qwen_image_series + wan_series + flux_series + flux2_series + z_image_series + ltx2_series + anima_series + mova_series
+stable_diffusion_xl_series = [
+    {
+        # Example: ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="unet/diffusion_pytorch_model.safetensors")
+        "model_hash": "142b114f67f5ab3a6d83fb5788f12ded",
+        "model_name": "stable_diffusion_xl_unet",
+        "model_class": "diffsynth.models.stable_diffusion_xl_unet.SDXLUNet2DConditionModel",
+        "extra_kwargs": {"attention_head_dim": [5, 10, 20], "transformer_layers_per_block": [1, 2, 10], "use_linear_projection": True, "addition_embed_type": "text_time", "addition_time_embed_dim": 256, "projection_class_embeddings_input_dim": 2816},
+    },
+    {
+        # Example: ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder_2/model.safetensors")
+        "model_hash": "98cc34ccc5b54ae0e56bdea8688dcd5a",
+        "model_name": "stable_diffusion_xl_text_encoder",
+        "model_class": "diffsynth.models.stable_diffusion_xl_text_encoder.SDXLTextEncoder2",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.stable_diffusion_xl_text_encoder.SDXLTextEncoder2StateDictConverter",
+    },
+    {
+        # Example: ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder/model.safetensors")
+        "model_hash": "94eefa3dac9cec93cb1ebaf1747d7b78",
+        "model_name": "stable_diffusion_text_encoder",
+        "model_class": "diffsynth.models.stable_diffusion_text_encoder.SDTextEncoder",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.stable_diffusion_text_encoder.SDTextEncoderStateDictConverter",
+    },
+    {
+        # Example: ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="vae/diffusion_pytorch_model.safetensors")
+        "model_hash": "13115dd45a6e1c39860f91ab073b8a78",
+        "model_name": "stable_diffusion_xl_vae",
+        "model_class": "diffsynth.models.stable_diffusion_vae.StableDiffusionVAE",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.stable_diffusion_vae.SDVAEStateDictConverter",
+        "extra_kwargs": {"scaling_factor": 0.13025, "sample_size": 1024, "force_upcast": True},
+    },
+]
+
+stable_diffusion_series = [
+    {
+        # Example: ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="text_encoder/model.safetensors")
+        "model_hash": "ffd1737ae9df7fd43f5fbed653bdad67",
+        "model_name": "stable_diffusion_text_encoder",
+        "model_class": "diffsynth.models.stable_diffusion_text_encoder.SDTextEncoder",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.stable_diffusion_text_encoder.SDTextEncoderStateDictConverter",
+    },
+    {
+        # Example: ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="vae/diffusion_pytorch_model.safetensors")
+        "model_hash": "f86d5683ed32433be8ca69969c67ba69",
+        "model_name": "stable_diffusion_vae",
+        "model_class": "diffsynth.models.stable_diffusion_vae.StableDiffusionVAE",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.stable_diffusion_vae.SDVAEStateDictConverter",
+    },
+    {
+        # Example: ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="unet/diffusion_pytorch_model.safetensors")
+        "model_hash": "025a4b86a84829399d89f613e580757b",
+        "model_name": "stable_diffusion_unet",
+        "model_class": "diffsynth.models.stable_diffusion_unet.UNet2DConditionModel",
+    },
+]
+
+joyai_image_series = [
+    {
+        # Example: ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="transformer/transformer.pth")
+        "model_hash": "56592ddfd7d0249d3aa527d24161a863",
+        "model_name": "joyai_image_dit",
+        "model_class": "diffsynth.models.joyai_image_dit.JoyAIImageDiT",
+    },
+    {
+        # Example: ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/model-*.safetensors")
+        "model_hash": "2d11bf14bba8b4e87477c8199a895403",
+        "model_name": "joyai_image_text_encoder",
+        "model_class": "diffsynth.models.joyai_image_text_encoder.JoyAIImageTextEncoder",
+        "state_dict_converter": "diffsynth.utils.state_dict_converters.joyai_image_text_encoder.JoyAIImageTextEncoderStateDictConverter",
+    },
+]
+
+MODEL_CONFIGS = stable_diffusion_xl_series + stable_diffusion_series + qwen_image_series + wan_series + flux_series + flux2_series + ernie_image_series + z_image_series + ltx2_series + anima_series + mova_series + joyai_image_series

diffsynth/configs/vram_management_module_maps.py

@@ -267,6 +267,73 @@ VRAM_MANAGEMENT_MODULE_MAPS = {
         "torch.nn.Conv1d": "diffsynth.core.vram.layers.AutoWrappedModule",
         "torch.nn.ConvTranspose1d": "diffsynth.core.vram.layers.AutoWrappedModule",
     },
+    "diffsynth.models.ernie_image_dit.ErnieImageDiT": {
+        "diffsynth.models.ernie_image_dit.ErnieImageRMSNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Conv2d": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.LayerNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.RMSNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.ernie_image_text_encoder.ErnieImageTextEncoder": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Embedding": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.ministral3.modeling_ministral3.Ministral3RMSNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.joyai_image_dit.Transformer3DModel": {
+        "diffsynth.models.joyai_image_dit.RMSNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "diffsynth.models.joyai_image_dit.ModulateWan": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Conv3d": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.LayerNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.joyai_image_text_encoder.JoyAIImageTextEncoder": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Embedding": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.LayerNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.Conv3d": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.qwen3_vl.modeling_qwen3_vl.Qwen3VLVisionModel": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.qwen3_vl.modeling_qwen3_vl.Qwen3VLTextRMSNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.qwen3_vl.modeling_qwen3_vl.Qwen3VLTextRotaryEmbedding": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.stable_diffusion_unet.UNet2DConditionModel": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Conv2d": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.GroupNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.LayerNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.SiLU": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.Dropout": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.stable_diffusion_vae.StableDiffusionVAE": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Conv2d": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.GroupNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.SiLU": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.Dropout": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "diffsynth.models.stable_diffusion_vae.Upsample2D": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "diffsynth.models.stable_diffusion_vae.Downsample2D": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.stable_diffusion_text_encoder.SDTextEncoder": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Embedding": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.clip.modeling_clip.CLIPTextTransformer": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.clip.modeling_clip.CLIPEncoderLayer": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.clip.modeling_clip.CLIPAttention": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.stable_diffusion_xl_unet.SDXLUNet2DConditionModel": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Conv2d": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.GroupNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.LayerNorm": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.SiLU": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "torch.nn.Dropout": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
+    "diffsynth.models.stable_diffusion_xl_text_encoder.SDXLTextEncoder2": {
+        "torch.nn.Linear": "diffsynth.core.vram.layers.AutoWrappedLinear",
+        "torch.nn.Embedding": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.clip.modeling_clip.CLIPTextTransformer": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.clip.modeling_clip.CLIPEncoderLayer": "diffsynth.core.vram.layers.AutoWrappedModule",
+        "transformers.models.clip.modeling_clip.CLIPAttention": "diffsynth.core.vram.layers.AutoWrappedModule",
+    },
 }
 
 def QwenImageTextEncoder_Module_Map_Updater():

diffsynth/core/data/operators.py

@@ -1,4 +1,4 @@
-import math
+import math, warnings
 import torch, torchvision, imageio, os
 import imageio.v3 as iio
 from PIL import Image
@@ -260,15 +260,19 @@ class LoadAudioWithTorchaudio(DataProcessingOperator, FrameSamplerByRateMixin):
         FrameSamplerByRateMixin.__init__(self, num_frames, time_division_factor, time_division_remainder, frame_rate, fix_frame_rate)
 
     def __call__(self, data: str):
-        reader = self.get_reader(data)
-        num_frames = self.get_num_frames(reader)
-        duration = num_frames / self.frame_rate
-        waveform, sample_rate = torchaudio.load(data)
-        target_samples = int(duration * sample_rate)
-        current_samples = waveform.shape[-1]
-        if current_samples > target_samples:
-            waveform = waveform[..., :target_samples]
-        elif current_samples < target_samples:
-            padding = target_samples - current_samples
-            waveform = torch.nn.functional.pad(waveform, (0, padding))
-        return waveform, sample_rate
+        try:
+            reader = self.get_reader(data)
+            num_frames = self.get_num_frames(reader)
+            duration = num_frames / self.frame_rate
+            waveform, sample_rate = torchaudio.load(data)
+            target_samples = int(duration * sample_rate)
+            current_samples = waveform.shape[-1]
+            if current_samples > target_samples:
+                waveform = waveform[..., :target_samples]
+            elif current_samples < target_samples:
+                padding = target_samples - current_samples
+                waveform = torch.nn.functional.pad(waveform, (0, padding))
+            return waveform, sample_rate
+        except:
+            warnings.warn(f"Cannot load audio in {data}. The audio will be `None`.")
+            return None

diffsynth/core/gradient/gradient_checkpoint.py

@@ -1,12 +1,32 @@
 import torch
 
 
+try:
+    import deepspeed
+    _HAS_DEEPSPEED = True
+except ModuleNotFoundError:
+    _HAS_DEEPSPEED = False
+
+
 def create_custom_forward(module):
     def custom_forward(*inputs, **kwargs):
         return module(*inputs, **kwargs)
     return custom_forward
 
 
+def create_custom_forward_use_reentrant(module):
+    def custom_forward(*inputs):
+        return module(*inputs)
+    return custom_forward
+
+
+def judge_args_requires_grad(*args):
+    for arg in args:
+        if isinstance(arg, torch.Tensor) and arg.requires_grad:
+            return True
+    return False
+
+
 def gradient_checkpoint_forward(
     model,
     use_gradient_checkpointing,
@@ -14,6 +34,17 @@ def gradient_checkpoint_forward(
     *args,
     **kwargs,
 ):
+    if use_gradient_checkpointing and _HAS_DEEPSPEED and deepspeed.checkpointing.is_configured():
+        all_args = args + tuple(kwargs.values())
+        if not judge_args_requires_grad(*all_args):
+            # get the first grad_enabled tensor from un_checkpointed forward
+            model_output = model(*args, **kwargs)
+        else:
+            model_output = deepspeed.checkpointing.checkpoint(
+                create_custom_forward_use_reentrant(model),
+                *all_args,
+            )
+        return model_output
     if use_gradient_checkpointing_offload:
         with torch.autograd.graph.save_on_cpu():
             model_output = torch.utils.checkpoint.checkpoint(

diffsynth/diffusion/base_pipeline.py

@@ -339,6 +339,38 @@ class BasePipeline(torch.nn.Module):
             noise_pred = noise_pred_posi
         return noise_pred
 
+    def compile_pipeline(self, mode: str = "default", dynamic: bool = True, fullgraph: bool = False, compile_models: list = None, **kwargs):
+        """
+        compile the pipeline with torch.compile. The models that will be compiled are determined by the `compilable_models` attribute of the pipeline.
+        If a model has `_repeated_blocks` attribute, we will compile these blocks with regional compilation. Otherwise, we will compile the whole model.
+        See https://docs.pytorch.org/docs/stable/generated/torch.compile.html#torch.compile for details about compilation arguments.
+        Args:
+            mode: The compilation mode, which will be passed to `torch.compile`, options are "default", "reduce-overhead", "max-autotune" and "max-autotune-no-cudagraphs. Default to "default".
+            dynamic: Whether to enable dynamic graph compilation to support dynamic input shapes, which will be passed to `torch.compile`. Default to True (recommended).
+            fullgraph: Whether to use full graph compilation, which will be passed to `torch.compile`. Default to False (recommended).
+            compile_models: The list of model names to be compiled. If None, we will compile the models in `pipeline.compilable_models`. Default to None.
+            **kwargs: Other arguments for `torch.compile`.
+        """
+        compile_models = compile_models or getattr(self, "compilable_models", [])
+        if len(compile_models) == 0:
+            print("No compilable models in the pipeline. Skip compilation.")
+            return
+        for name in compile_models:
+            model = getattr(self, name, None)
+            if model is None:
+                print(f"Model '{name}' not found in the pipeline.")
+                continue
+            repeated_blocks = getattr(model, "_repeated_blocks", None)
+            # regional compilation for repeated blocks.
+            if repeated_blocks is not None:
+                for submod in model.modules():
+                    if submod.__class__.__name__ in repeated_blocks:
+                        submod.compile(mode=mode, dynamic=dynamic, fullgraph=fullgraph, **kwargs)
+            # compile the whole model.
+            else:
+                model.compile(mode=mode, dynamic=dynamic, fullgraph=fullgraph, **kwargs)
+            print(f"{name} is compiled with mode={mode}, dynamic={dynamic}, fullgraph={fullgraph}.")
+
 
 class PipelineUnitGraph:
     def __init__(self):

diffsynth/diffusion/ddim_scheduler.py

@@ -0,0 +1,107 @@
+import torch, math
+
+
+class DDIMScheduler():
+
+    def __init__(self, num_train_timesteps=1000, beta_start=0.00085, beta_end=0.012, beta_schedule="scaled_linear", prediction_type="epsilon", rescale_zero_terminal_snr=False):
+        self.num_train_timesteps = num_train_timesteps
+        if beta_schedule == "scaled_linear":
+            betas = torch.square(torch.linspace(math.sqrt(beta_start), math.sqrt(beta_end), num_train_timesteps, dtype=torch.float32))
+        elif beta_schedule == "linear":
+            betas = torch.linspace(beta_start, beta_end, num_train_timesteps, dtype=torch.float32)
+        else:
+            raise NotImplementedError(f"{beta_schedule} is not implemented")
+        self.alphas_cumprod = torch.cumprod(1.0 - betas, dim=0)
+        if rescale_zero_terminal_snr:
+            self.alphas_cumprod = self.rescale_zero_terminal_snr(self.alphas_cumprod)
+        self.alphas_cumprod = self.alphas_cumprod.tolist()
+        self.set_timesteps(10)
+        self.prediction_type = prediction_type
+        self.training = False
+
+
+    def rescale_zero_terminal_snr(self, alphas_cumprod):
+        alphas_bar_sqrt = alphas_cumprod.sqrt()
+
+        # Store old values.
+        alphas_bar_sqrt_0 = alphas_bar_sqrt[0].clone()
+        alphas_bar_sqrt_T = alphas_bar_sqrt[-1].clone()
+
+        # Shift so the last timestep is zero.
+        alphas_bar_sqrt -= alphas_bar_sqrt_T
+
+        # Scale so the first timestep is back to the old value.
+        alphas_bar_sqrt *= alphas_bar_sqrt_0 / (alphas_bar_sqrt_0 - alphas_bar_sqrt_T)
+
+        # Convert alphas_bar_sqrt to betas
+        alphas_bar = alphas_bar_sqrt.square()  # Revert sqrt
+
+        return alphas_bar
+
+
+    def set_timesteps(self, num_inference_steps, denoising_strength=1.0, training=False, **kwargs):
+        # The timesteps are aligned to 999...0, which is different from other implementations,
+        # but I think this implementation is more reasonable in theory.
+        max_timestep = max(round(self.num_train_timesteps * denoising_strength) - 1, 0)
+        num_inference_steps = min(num_inference_steps, max_timestep + 1)
+        if num_inference_steps == 1:
+            self.timesteps = torch.Tensor([max_timestep])
+        else:
+            step_length = max_timestep / (num_inference_steps - 1)
+            self.timesteps = torch.Tensor([round(max_timestep - i*step_length) for i in range(num_inference_steps)])
+        self.training = training
+
+
+    def denoise(self, model_output, sample, alpha_prod_t, alpha_prod_t_prev):
+        if self.prediction_type == "epsilon":
+            weight_e = math.sqrt(1 - alpha_prod_t_prev) - math.sqrt(alpha_prod_t_prev * (1 - alpha_prod_t) / alpha_prod_t)
+            weight_x = math.sqrt(alpha_prod_t_prev / alpha_prod_t)
+            prev_sample = sample * weight_x + model_output * weight_e
+        elif self.prediction_type == "v_prediction":
+            weight_e = -math.sqrt(alpha_prod_t_prev * (1 - alpha_prod_t)) + math.sqrt(alpha_prod_t * (1 - alpha_prod_t_prev))
+            weight_x = math.sqrt(alpha_prod_t * alpha_prod_t_prev) + math.sqrt((1 - alpha_prod_t) * (1 - alpha_prod_t_prev))
+            prev_sample = sample * weight_x + model_output * weight_e
+        else:
+            raise NotImplementedError(f"{self.prediction_type} is not implemented")
+        return prev_sample
+
+
+    def step(self, model_output, timestep, sample, to_final=False):
+        alpha_prod_t = self.alphas_cumprod[int(timestep.flatten().tolist()[0])]
+        if isinstance(timestep, torch.Tensor):
+            timestep = timestep.cpu()
+        timestep_id = torch.argmin((self.timesteps - timestep).abs())
+        if to_final or timestep_id + 1 >= len(self.timesteps):
+            alpha_prod_t_prev = 1.0
+        else:
+            timestep_prev = int(self.timesteps[timestep_id + 1])
+            alpha_prod_t_prev = self.alphas_cumprod[timestep_prev]
+
+        return self.denoise(model_output, sample, alpha_prod_t, alpha_prod_t_prev)
+
+
+    def return_to_timestep(self, timestep, sample, sample_stablized):
+        alpha_prod_t = self.alphas_cumprod[int(timestep.flatten().tolist()[0])]
+        noise_pred = (sample - math.sqrt(alpha_prod_t) * sample_stablized) / math.sqrt(1 - alpha_prod_t)
+        return noise_pred
+    
+    
+    def add_noise(self, original_samples, noise, timestep):
+        sqrt_alpha_prod = math.sqrt(self.alphas_cumprod[int(timestep.flatten().tolist()[0])])
+        sqrt_one_minus_alpha_prod = math.sqrt(1 - self.alphas_cumprod[int(timestep.flatten().tolist()[0])])
+        noisy_samples = sqrt_alpha_prod * original_samples + sqrt_one_minus_alpha_prod * noise
+        return noisy_samples
+    
+
+    def training_target(self, sample, noise, timestep):
+        if self.prediction_type == "epsilon":
+            return noise
+        else:
+            sqrt_alpha_prod = math.sqrt(self.alphas_cumprod[int(timestep.flatten().tolist()[0])])
+            sqrt_one_minus_alpha_prod = math.sqrt(1 - self.alphas_cumprod[int(timestep.flatten().tolist()[0])])
+            target = sqrt_alpha_prod * noise - sqrt_one_minus_alpha_prod * sample
+            return target
+        
+    
+    def training_weight(self, timestep):
+        return 1.0

diffsynth/diffusion/flow_match.py

@@ -4,7 +4,7 @@ from typing_extensions import Literal
 
 class FlowMatchScheduler():
 
-    def __init__(self, template: Literal["FLUX.1", "Wan", "Qwen-Image", "FLUX.2", "Z-Image", "LTX-2", "Qwen-Image-Lightning"] = "FLUX.1"):
+    def __init__(self, template: Literal["FLUX.1", "Wan", "Qwen-Image", "FLUX.2", "Z-Image", "LTX-2", "Qwen-Image-Lightning", "ERNIE-Image"] = "FLUX.1"):
         self.set_timesteps_fn = {
             "FLUX.1": FlowMatchScheduler.set_timesteps_flux,
             "Wan": FlowMatchScheduler.set_timesteps_wan,
@@ -13,6 +13,7 @@ class FlowMatchScheduler():
             "Z-Image": FlowMatchScheduler.set_timesteps_z_image,
             "LTX-2": FlowMatchScheduler.set_timesteps_ltx2,
             "Qwen-Image-Lightning": FlowMatchScheduler.set_timesteps_qwen_image_lightning,
+            "ERNIE-Image": FlowMatchScheduler.set_timesteps_ernie_image,
         }.get(template, FlowMatchScheduler.set_timesteps_flux)
         self.num_train_timesteps = 1000
 
@@ -129,6 +130,18 @@ class FlowMatchScheduler():
         timesteps = sigmas * num_train_timesteps
         return sigmas, timesteps
 
+    @staticmethod
+    def set_timesteps_ernie_image(num_inference_steps=50, denoising_strength=1.0, shift=3.0):
+        sigma_min = 0.0
+        sigma_max = 1.0
+        num_train_timesteps = 1000
+        sigma_start = sigma_min + (sigma_max - sigma_min) * denoising_strength
+        sigmas = torch.linspace(sigma_start, sigma_min, num_inference_steps + 1)[:-1]
+        if shift is not None and shift != 1.0:
+            sigmas = shift * sigmas / (1 + (shift - 1) * sigmas)
+        timesteps = sigmas * num_train_timesteps
+        return sigmas, timesteps
+
     @staticmethod
     def set_timesteps_z_image(num_inference_steps=100, denoising_strength=1.0, shift=None, target_timesteps=None):
         sigma_min = 0.0
@@ -146,6 +159,18 @@ class FlowMatchScheduler():
                 timesteps[timestep_id] = timestep
         return sigmas, timesteps
 
+    @staticmethod
+    def set_timesteps_joyai_image(num_inference_steps=100, denoising_strength=1.0, shift=None):
+        sigma_min = 0.0
+        sigma_max = 1.0
+        shift = 4.0 if shift is None else shift
+        num_train_timesteps = 1000
+        sigma_start = sigma_min + (sigma_max - sigma_min) * denoising_strength
+        sigmas = torch.linspace(sigma_start, sigma_min, num_inference_steps + 1)[:-1]
+        sigmas = shift * sigmas / (1 + (shift - 1) * sigmas)
+        timesteps = sigmas * num_train_timesteps
+        return sigmas, timesteps
+
     @staticmethod
     def set_timesteps_ltx2(num_inference_steps=100, denoising_strength=1.0, dynamic_shift_len=None, terminal=0.1, special_case=None):
         num_train_timesteps = 1000
@@ -185,7 +210,7 @@ class FlowMatchScheduler():
             bsmntw_weighing = bsmntw_weighing * (len(self.timesteps) / steps)
             bsmntw_weighing = bsmntw_weighing + bsmntw_weighing[1]
         self.linear_timesteps_weights = bsmntw_weighing
-        
+
     def set_timesteps(self, num_inference_steps=100, denoising_strength=1.0, training=False, **kwargs):
         self.sigmas, self.timesteps = self.set_timesteps_fn(
             num_inference_steps=num_inference_steps,

diffsynth/diffusion/runner.py

@@ -29,19 +29,19 @@ def launch_training_task(
     dataloader = torch.utils.data.DataLoader(dataset, shuffle=True, collate_fn=lambda x: x[0], num_workers=num_workers)
     model.to(device=accelerator.device)
     model, optimizer, dataloader, scheduler = accelerator.prepare(model, optimizer, dataloader, scheduler)
-    
+    initialize_deepspeed_gradient_checkpointing(accelerator)
     for epoch_id in range(num_epochs):
         for data in tqdm(dataloader):
             with accelerator.accumulate(model):
-                optimizer.zero_grad()
                 if dataset.load_from_cache:
                     loss = model({}, inputs=data)
                 else:
                     loss = model(data)
                 accelerator.backward(loss)
                 optimizer.step()
-                model_logger.on_step_end(accelerator, model, save_steps, loss=loss)
                 scheduler.step()
+                optimizer.zero_grad()
+                model_logger.on_step_end(accelerator, model, save_steps, loss=loss)
         if save_steps is None:
             model_logger.on_epoch_end(accelerator, model, epoch_id)
     model_logger.on_training_end(accelerator, model, save_steps)
@@ -70,3 +70,19 @@ def launch_data_process_task(
                 save_path = os.path.join(model_logger.output_path, str(accelerator.process_index), f"{data_id}.pth")
                 data = model(data)
                 torch.save(data, save_path)
+
+
+def initialize_deepspeed_gradient_checkpointing(accelerator: Accelerator):
+    if getattr(accelerator.state, "deepspeed_plugin", None) is not None:
+        ds_config = accelerator.state.deepspeed_plugin.deepspeed_config
+        if "activation_checkpointing" in ds_config:
+            import deepspeed
+            act_config = ds_config["activation_checkpointing"]
+            deepspeed.checkpointing.configure(
+                mpu_=None, 
+                partition_activations=act_config.get("partition_activations", False),
+                checkpoint_in_cpu=act_config.get("cpu_checkpointing", False),
+                contiguous_checkpointing=act_config.get("contiguous_memory_optimization", False)
+            )
+        else:
+            print("Do not find activation_checkpointing config in deepspeed config, skip initializing deepspeed gradient checkpointing.")

diffsynth/models/anima_dit.py

@@ -1270,6 +1270,9 @@ class LLMAdapter(nn.Module):
 
 
 class AnimaDiT(MiniTrainDIT):
+
+    _repeated_blocks = ["Block"]
+
     def __init__(self):
         kwargs = {'image_model': 'anima', 'max_img_h': 240, 'max_img_w': 240, 'max_frames': 128, 'in_channels': 16, 'out_channels': 16, 'patch_spatial': 2, 'patch_temporal': 1, 'model_channels': 2048, 'concat_padding_mask': True, 'crossattn_emb_channels': 1024, 'pos_emb_cls': 'rope3d', 'pos_emb_learnable': True, 'pos_emb_interpolation': 'crop', 'min_fps': 1, 'max_fps': 30, 'use_adaln_lora': True, 'adaln_lora_dim': 256, 'num_blocks': 28, 'num_heads': 16, 'extra_per_block_abs_pos_emb': False, 'rope_h_extrapolation_ratio': 4.0, 'rope_w_extrapolation_ratio': 4.0, 'rope_t_extrapolation_ratio': 1.0, 'extra_h_extrapolation_ratio': 1.0, 'extra_w_extrapolation_ratio': 1.0, 'extra_t_extrapolation_ratio': 1.0, 'rope_enable_fps_modulation': False, 'dtype': torch.bfloat16, 'device': None, 'operations': torch.nn}
         super().__init__(**kwargs)

diffsynth/models/ernie_image_dit.py

@@ -0,0 +1,362 @@
+"""
+Ernie-Image DiT for DiffSynth-Studio.
+
+Refactored from diffusers ErnieImageTransformer2DModel to use DiffSynth core modules.
+Default parameters from actual checkpoint config.json (PaddlePaddle/ERNIE-Image transformer).
+"""
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from typing import Optional, Tuple
+
+from ..core.attention import attention_forward
+from ..core.gradient import gradient_checkpoint_forward
+from .flux2_dit import Timesteps, TimestepEmbedding
+
+
+def rope(pos: torch.Tensor, dim: int, theta: int) -> torch.Tensor:
+    assert dim % 2 == 0
+    scale = torch.arange(0, dim, 2, dtype=torch.float64, device=pos.device) / dim
+    omega = 1.0 / (theta ** scale)
+    out = torch.einsum("...n,d->...nd", pos, omega)
+    return out.float()
+
+
+class ErnieImageEmbedND3(nn.Module):
+    def __init__(self, dim: int, theta: int, axes_dim: Tuple[int, int, int]):
+        super().__init__()
+        self.dim = dim
+        self.theta = theta
+        self.axes_dim = list(axes_dim)
+
+    def forward(self, ids: torch.Tensor) -> torch.Tensor:
+        emb = torch.cat([rope(ids[..., i], self.axes_dim[i], self.theta) for i in range(3)], dim=-1)
+        emb = emb.unsqueeze(2)
+        return torch.stack([emb, emb], dim=-1).reshape(*emb.shape[:-1], -1)
+
+
+class ErnieImagePatchEmbedDynamic(nn.Module):
+    def __init__(self, in_channels: int, embed_dim: int, patch_size: int):
+        super().__init__()
+        self.patch_size = patch_size
+        self.proj = nn.Conv2d(in_channels, embed_dim, kernel_size=patch_size, stride=patch_size, bias=True)
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        x = self.proj(x)
+        batch_size, dim, height, width = x.shape
+        return x.reshape(batch_size, dim, height * width).transpose(1, 2).contiguous()
+
+
+class ErnieImageSingleStreamAttnProcessor:
+    def __call__(
+        self,
+        attn: "ErnieImageAttention",
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        freqs_cis: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        query = attn.to_q(hidden_states)
+        key = attn.to_k(hidden_states)
+        value = attn.to_v(hidden_states)
+
+        query = query.unflatten(-1, (attn.heads, -1))
+        key = key.unflatten(-1, (attn.heads, -1))
+        value = value.unflatten(-1, (attn.heads, -1))
+
+        if attn.norm_q is not None:
+            query = attn.norm_q(query)
+        if attn.norm_k is not None:
+            key = attn.norm_k(key)
+
+        def apply_rotary_emb(x_in: torch.Tensor, freqs_cis: torch.Tensor) -> torch.Tensor:
+            rot_dim = freqs_cis.shape[-1]
+            x, x_pass = x_in[..., :rot_dim], x_in[..., rot_dim:]
+            cos_ = torch.cos(freqs_cis).to(x.dtype)
+            sin_ = torch.sin(freqs_cis).to(x.dtype)
+            x1, x2 = x.chunk(2, dim=-1)
+            x_rotated = torch.cat((-x2, x1), dim=-1)
+            return torch.cat((x * cos_ + x_rotated * sin_, x_pass), dim=-1)
+
+        if freqs_cis is not None:
+            query = apply_rotary_emb(query, freqs_cis)
+            key = apply_rotary_emb(key, freqs_cis)
+
+        if attention_mask is not None and attention_mask.ndim == 2:
+            attention_mask = attention_mask[:, None, None, :]
+
+        hidden_states = attention_forward(
+            query, key, value,
+            q_pattern="b s n d",
+            k_pattern="b s n d",
+            v_pattern="b s n d",
+            out_pattern="b s n d",
+            attn_mask=attention_mask,
+        )
+
+        hidden_states = hidden_states.flatten(2, 3)
+        hidden_states = hidden_states.to(query.dtype)
+        output = attn.to_out[0](hidden_states)
+
+        return output
+
+
+class ErnieImageAttention(nn.Module):
+    def __init__(
+        self,
+        query_dim: int,
+        heads: int = 8,
+        dim_head: int = 64,
+        dropout: float = 0.0,
+        bias: bool = False,
+        qk_norm: str = "rms_norm",
+        out_bias: bool = True,
+        eps: float = 1e-5,
+        out_dim: int = None,
+        elementwise_affine: bool = True,
+    ):
+        super().__init__()
+
+        self.head_dim = dim_head
+        self.inner_dim = out_dim if out_dim is not None else dim_head * heads
+        self.query_dim = query_dim
+        self.out_dim = out_dim if out_dim is not None else query_dim
+        self.heads = out_dim // dim_head if out_dim is not None else heads
+
+        self.use_bias = bias
+        self.dropout = dropout
+
+        self.to_q = nn.Linear(query_dim, self.inner_dim, bias=bias)
+        self.to_k = nn.Linear(query_dim, self.inner_dim, bias=bias)
+        self.to_v = nn.Linear(query_dim, self.inner_dim, bias=bias)
+
+        if qk_norm == "layer_norm":
+            self.norm_q = nn.LayerNorm(dim_head, eps=eps, elementwise_affine=elementwise_affine)
+            self.norm_k = nn.LayerNorm(dim_head, eps=eps, elementwise_affine=elementwise_affine)
+        elif qk_norm == "rms_norm":
+            self.norm_q = nn.RMSNorm(dim_head, eps=eps, elementwise_affine=elementwise_affine)
+            self.norm_k = nn.RMSNorm(dim_head, eps=eps, elementwise_affine=elementwise_affine)
+        else:
+            raise ValueError(
+                f"unknown qk_norm: {qk_norm}. Should be one of None, 'layer_norm', 'rms_norm'."
+            )
+
+        self.to_out = nn.ModuleList([])
+        self.to_out.append(nn.Linear(self.inner_dim, self.out_dim, bias=out_bias))
+
+        self.processor = ErnieImageSingleStreamAttnProcessor()
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        image_rotary_emb: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        return self.processor(self, hidden_states, attention_mask, image_rotary_emb)
+
+
+class ErnieImageFeedForward(nn.Module):
+    def __init__(self, hidden_size: int, ffn_hidden_size: int):
+        super().__init__()
+        self.gate_proj = nn.Linear(hidden_size, ffn_hidden_size, bias=False)
+        self.up_proj = nn.Linear(hidden_size, ffn_hidden_size, bias=False)
+        self.linear_fc2 = nn.Linear(ffn_hidden_size, hidden_size, bias=False)
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.linear_fc2(self.up_proj(x) * F.gelu(self.gate_proj(x)))
+
+
+class ErnieImageRMSNorm(nn.Module):
+    def __init__(self, dim: int, eps: float = 1e-6):
+        super().__init__()
+        self.eps = eps
+        self.weight = nn.Parameter(torch.ones(dim))
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        input_dtype = hidden_states.dtype
+        variance = hidden_states.to(torch.float32).pow(2).mean(-1, keepdim=True)
+        hidden_states = hidden_states * torch.rsqrt(variance + self.eps)
+        hidden_states = hidden_states * self.weight
+        return hidden_states.to(input_dtype)
+
+
+class ErnieImageSharedAdaLNBlock(nn.Module):
+    def __init__(
+        self,
+        hidden_size: int,
+        num_heads: int,
+        ffn_hidden_size: int,
+        eps: float = 1e-6,
+        qk_layernorm: bool = True,
+    ):
+        super().__init__()
+        self.adaLN_sa_ln = ErnieImageRMSNorm(hidden_size, eps=eps)
+        self.self_attention = ErnieImageAttention(
+            query_dim=hidden_size,
+            dim_head=hidden_size // num_heads,
+            heads=num_heads,
+            qk_norm="rms_norm" if qk_layernorm else None,
+            eps=eps,
+            bias=False,
+            out_bias=False,
+        )
+        self.adaLN_mlp_ln = ErnieImageRMSNorm(hidden_size, eps=eps)
+        self.mlp = ErnieImageFeedForward(hidden_size, ffn_hidden_size)
+
+    def forward(
+        self,
+        x: torch.Tensor,
+        rotary_pos_emb: torch.Tensor,
+        temb: Tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor],
+        attention_mask: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp = temb
+        residual = x
+        x = self.adaLN_sa_ln(x)
+        x = (x.float() * (1 + scale_msa.float()) + shift_msa.float()).to(x.dtype)
+        x_bsh = x.permute(1, 0, 2)
+        attn_out = self.self_attention(x_bsh, attention_mask=attention_mask, image_rotary_emb=rotary_pos_emb)
+        attn_out = attn_out.permute(1, 0, 2)
+        x = residual + (gate_msa.float() * attn_out.float()).to(x.dtype)
+        residual = x
+        x = self.adaLN_mlp_ln(x)
+        x = (x.float() * (1 + scale_mlp.float()) + shift_mlp.float()).to(x.dtype)
+        return residual + (gate_mlp.float() * self.mlp(x).float()).to(x.dtype)
+
+
+class ErnieImageAdaLNContinuous(nn.Module):
+    def __init__(self, hidden_size: int, eps: float = 1e-6):
+        super().__init__()
+        self.norm = nn.LayerNorm(hidden_size, elementwise_affine=False, eps=eps)
+        self.linear = nn.Linear(hidden_size, hidden_size * 2)
+
+    def forward(self, x: torch.Tensor, conditioning: torch.Tensor) -> torch.Tensor:
+        scale, shift = self.linear(conditioning).chunk(2, dim=-1)
+        x = self.norm(x)
+        x = x * (1 + scale.unsqueeze(0)) + shift.unsqueeze(0)
+        return x
+
+
+class ErnieImageDiT(nn.Module):
+    """
+    Ernie-Image DiT model for DiffSynth-Studio.
+
+    Architecture: SharedAdaLN + RoPE 3D + Joint Image-Text Attention.
+    Internal format: [S, B, H] for transformer blocks, [B, S, H] for attention.
+    """
+
+    def __init__(
+        self,
+        hidden_size: int = 4096,
+        num_attention_heads: int = 32,
+        num_layers: int = 36,
+        ffn_hidden_size: int = 12288,
+        in_channels: int = 128,
+        out_channels: int = 128,
+        patch_size: int = 1,
+        text_in_dim: int = 3072,
+        rope_theta: int = 256,
+        rope_axes_dim: Tuple[int, int, int] = (32, 48, 48),
+        eps: float = 1e-6,
+        qk_layernorm: bool = True,
+    ):
+        super().__init__()
+        self.hidden_size = hidden_size
+        self.num_heads = num_attention_heads
+        self.head_dim = hidden_size // num_attention_heads
+        self.num_layers = num_layers
+        self.patch_size = patch_size
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.text_in_dim = text_in_dim
+
+        self.x_embedder = ErnieImagePatchEmbedDynamic(in_channels, hidden_size, patch_size)
+        self.text_proj = nn.Linear(text_in_dim, hidden_size, bias=False) if text_in_dim != hidden_size else None
+        self.time_proj = Timesteps(hidden_size, flip_sin_to_cos=False, downscale_freq_shift=0)
+        self.time_embedding = TimestepEmbedding(hidden_size, hidden_size)
+        self.pos_embed = ErnieImageEmbedND3(dim=self.head_dim, theta=rope_theta, axes_dim=rope_axes_dim)
+        self.adaLN_modulation = nn.Sequential(nn.SiLU(), nn.Linear(hidden_size, 6 * hidden_size))
+        nn.init.zeros_(self.adaLN_modulation[-1].weight)
+        nn.init.zeros_(self.adaLN_modulation[-1].bias)
+        self.layers = nn.ModuleList([
+            ErnieImageSharedAdaLNBlock(hidden_size, num_attention_heads, ffn_hidden_size, eps, qk_layernorm=qk_layernorm)
+            for _ in range(num_layers)
+        ])
+        self.final_norm = ErnieImageAdaLNContinuous(hidden_size, eps)
+        self.final_linear = nn.Linear(hidden_size, patch_size * patch_size * out_channels)
+        nn.init.zeros_(self.final_linear.weight)
+        nn.init.zeros_(self.final_linear.bias)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        timestep: torch.Tensor,
+        text_bth: torch.Tensor,
+        text_lens: torch.Tensor,
+        use_gradient_checkpointing: bool = False,
+        use_gradient_checkpointing_offload: bool = False,
+    ) -> torch.Tensor:
+        device, dtype = hidden_states.device, hidden_states.dtype
+        B, C, H, W = hidden_states.shape
+        p, Hp, Wp = self.patch_size, H // self.patch_size, W // self.patch_size
+        N_img = Hp * Wp
+
+        img_sbh = self.x_embedder(hidden_states).transpose(0, 1).contiguous()
+
+        if self.text_proj is not None and text_bth.numel() > 0:
+            text_bth = self.text_proj(text_bth)
+        Tmax = text_bth.shape[1]
+        text_sbh = text_bth.transpose(0, 1).contiguous()
+
+        x = torch.cat([img_sbh, text_sbh], dim=0)
+        S = x.shape[0]
+
+        text_ids = torch.cat([
+            torch.arange(Tmax, device=device, dtype=torch.float32).view(1, Tmax, 1).expand(B, -1, -1),
+            torch.zeros((B, Tmax, 2), device=device)
+        ], dim=-1) if Tmax > 0 else torch.zeros((B, 0, 3), device=device)
+        grid_yx = torch.stack(
+            torch.meshgrid(torch.arange(Hp, device=device, dtype=torch.float32),
+                           torch.arange(Wp, device=device, dtype=torch.float32), indexing="ij"),
+            dim=-1
+        ).reshape(-1, 2)
+        image_ids = torch.cat([
+            text_lens.float().view(B, 1, 1).expand(-1, N_img, -1),
+            grid_yx.view(1, N_img, 2).expand(B, -1, -1)
+        ], dim=-1)
+        rotary_pos_emb = self.pos_embed(torch.cat([image_ids, text_ids], dim=1))
+
+        valid_text = torch.arange(Tmax, device=device).view(1, Tmax) < text_lens.view(B, 1) if Tmax > 0 else torch.zeros((B, 0), device=device, dtype=torch.bool)
+        attention_mask = torch.cat([
+            torch.ones((B, N_img), device=device, dtype=torch.bool),
+            valid_text
+        ], dim=1)[:, None, None, :]
+
+        sample = self.time_proj(timestep.to(dtype))
+        sample = sample.to(self.time_embedding.linear_1.weight.dtype)
+        c = self.time_embedding(sample)
+        shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp = [
+            t.unsqueeze(0).expand(S, -1, -1).contiguous()
+            for t in self.adaLN_modulation(c).chunk(6, dim=-1)
+        ]
+
+        for layer in self.layers:
+            temb = [shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp]
+            if torch.is_grad_enabled() and use_gradient_checkpointing:
+                x = gradient_checkpoint_forward(
+                    layer,
+                    use_gradient_checkpointing,
+                    use_gradient_checkpointing_offload,
+                    x,
+                    rotary_pos_emb,
+                    temb,
+                    attention_mask,
+                )
+            else:
+                x = layer(x, rotary_pos_emb, temb, attention_mask)
+
+        x = self.final_norm(x, c).type_as(x)
+        patches = self.final_linear(x)[:N_img].transpose(0, 1).contiguous()
+        output = patches.view(B, Hp, Wp, p, p, self.out_channels).permute(0, 5, 1, 3, 2, 4).contiguous().view(B, self.out_channels, H, W)
+
+        return output

diffsynth/models/ernie_image_text_encoder.py

@@ -0,0 +1,76 @@
+"""
+Ernie-Image TextEncoder for DiffSynth-Studio.
+
+Wraps transformers Ministral3Model to output text embeddings.
+Pattern: lazy import + manual config dict + torch.nn.Module wrapper.
+Only loads the text (language) model, ignoring vision components.
+"""
+
+import torch
+
+
+class ErnieImageTextEncoder(torch.nn.Module):
+    """
+    Text encoder using Ministral3Model (transformers).
+    Only the text_config portion of the full Mistral3Model checkpoint.
+    Uses the base model (no lm_head) since the checkpoint only has embeddings.
+    """
+
+    def __init__(self):
+        super().__init__()
+        from transformers import Ministral3Config, Ministral3Model
+
+        text_config = {
+            "attention_dropout": 0.0,
+            "bos_token_id": 1,
+            "dtype": "bfloat16",
+            "eos_token_id": 2,
+            "head_dim": 128,
+            "hidden_act": "silu",
+            "hidden_size": 3072,
+            "initializer_range": 0.02,
+            "intermediate_size": 9216,
+            "max_position_embeddings": 262144,
+            "model_type": "ministral3",
+            "num_attention_heads": 32,
+            "num_hidden_layers": 26,
+            "num_key_value_heads": 8,
+            "pad_token_id": 11,
+            "rms_norm_eps": 1e-05,
+            "rope_parameters": {
+                "beta_fast": 32.0,
+                "beta_slow": 1.0,
+                "factor": 16.0,
+                "llama_4_scaling_beta": 0.1,
+                "mscale": 1.0,
+                "mscale_all_dim": 1.0,
+                "original_max_position_embeddings": 16384,
+                "rope_theta": 1000000.0,
+                "rope_type": "yarn",
+                "type": "yarn",
+            },
+            "sliding_window": None,
+            "tie_word_embeddings": True,
+            "use_cache": True,
+            "vocab_size": 131072,
+        }
+        config = Ministral3Config(**text_config)
+        self.model = Ministral3Model(config)
+        self.config = config
+
+    def forward(
+        self,
+        input_ids=None,
+        attention_mask=None,
+        position_ids=None,
+        **kwargs,
+    ):
+        outputs = self.model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            output_hidden_states=True,
+            return_dict=True,
+            **kwargs,
+        )
+        return (outputs.hidden_states,)

diffsynth/models/flux2_dit.py

@@ -879,6 +879,9 @@ class Flux2Modulation(nn.Module):
 
 
 class Flux2DiT(torch.nn.Module):
+
+    _repeated_blocks = ["Flux2TransformerBlock", "Flux2SingleTransformerBlock"]
+
     def __init__(
         self,
         patch_size: int = 1,

diffsynth/models/flux_dit.py

@@ -275,6 +275,9 @@ class AdaLayerNormContinuous(torch.nn.Module):
 
 
 class FluxDiT(torch.nn.Module):
+
+    _repeated_blocks = ["FluxJointTransformerBlock", "FluxSingleTransformerBlock"]
+
     def __init__(self, disable_guidance_embedder=False, input_dim=64, num_blocks=19):
         super().__init__()
         self.pos_embedder = RoPEEmbedding(3072, 10000, [16, 56, 56])

diffsynth/models/joyai_image_dit.py

@@ -0,0 +1,636 @@
+import math
+from typing import Dict, List, Optional, Tuple, Union
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from einops import rearrange
+
+from ..core.attention import attention_forward
+from ..core.gradient import gradient_checkpoint_forward
+
+
+def get_timestep_embedding(
+    timesteps: torch.Tensor,
+    embedding_dim: int,
+    flip_sin_to_cos: bool = False,
+    downscale_freq_shift: float = 1,
+    scale: float = 1,
+    max_period: int = 10000,
+) -> torch.Tensor:
+    assert len(timesteps.shape) == 1, "Timesteps should be a 1d-array"
+    half_dim = embedding_dim // 2
+    exponent = -math.log(max_period) * torch.arange(
+        start=0, end=half_dim, dtype=torch.float32, device=timesteps.device
+    )
+    exponent = exponent / (half_dim - downscale_freq_shift)
+    emb = torch.exp(exponent)
+    emb = timesteps[:, None].float() * emb[None, :]
+    emb = scale * emb
+    emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=-1)
+    if flip_sin_to_cos:
+        emb = torch.cat([emb[:, half_dim:], emb[:, :half_dim]], dim=-1)
+    if embedding_dim % 2 == 1:
+        emb = torch.nn.functional.pad(emb, (0, 1, 0, 0))
+    return emb
+
+
+class Timesteps(nn.Module):
+    def __init__(self, num_channels: int, flip_sin_to_cos: bool, downscale_freq_shift: float, scale: int = 1):
+        super().__init__()
+        self.num_channels = num_channels
+        self.flip_sin_to_cos = flip_sin_to_cos
+        self.downscale_freq_shift = downscale_freq_shift
+        self.scale = scale
+
+    def forward(self, timesteps: torch.Tensor) -> torch.Tensor:
+        return get_timestep_embedding(
+            timesteps,
+            self.num_channels,
+            flip_sin_to_cos=self.flip_sin_to_cos,
+            downscale_freq_shift=self.downscale_freq_shift,
+            scale=self.scale,
+        )
+
+
+class TimestepEmbedding(nn.Module):
+    def __init__(
+        self,
+        in_channels: int,
+        time_embed_dim: int,
+        act_fn: str = "silu",
+        out_dim: int = None,
+        post_act_fn: Optional[str] = None,
+        cond_proj_dim=None,
+        sample_proj_bias=True,
+    ):
+        super().__init__()
+        self.linear_1 = nn.Linear(in_channels, time_embed_dim, sample_proj_bias)
+        if cond_proj_dim is not None:
+            self.cond_proj = nn.Linear(cond_proj_dim, in_channels, bias=False)
+        else:
+            self.cond_proj = None
+        self.act = nn.SiLU()
+        time_embed_dim_out = out_dim if out_dim is not None else time_embed_dim
+        self.linear_2 = nn.Linear(time_embed_dim, time_embed_dim_out, sample_proj_bias)
+        self.post_act = nn.SiLU() if post_act_fn == "silu" else None
+
+    def forward(self, sample, condition=None):
+        if condition is not None:
+            sample = sample + self.cond_proj(condition)
+        sample = self.linear_1(sample)
+        if self.act is not None:
+            sample = self.act(sample)
+        sample = self.linear_2(sample)
+        if self.post_act is not None:
+            sample = self.post_act(sample)
+        return sample
+
+
+class PixArtAlphaTextProjection(nn.Module):
+    def __init__(self, in_features, hidden_size, out_features=None, act_fn="gelu_tanh"):
+        super().__init__()
+        if out_features is None:
+            out_features = hidden_size
+        self.linear_1 = nn.Linear(in_features=in_features, out_features=hidden_size, bias=True)
+        if act_fn == "gelu_tanh":
+            self.act_1 = nn.GELU(approximate="tanh")
+        elif act_fn == "silu":
+            self.act_1 = nn.SiLU()
+        else:
+            self.act_1 = nn.GELU(approximate="tanh")
+        self.linear_2 = nn.Linear(in_features=hidden_size, out_features=out_features, bias=True)
+
+    def forward(self, caption):
+        hidden_states = self.linear_1(caption)
+        hidden_states = self.act_1(hidden_states)
+        hidden_states = self.linear_2(hidden_states)
+        return hidden_states
+
+
+class GELU(nn.Module):
+    def __init__(self, dim_in: int, dim_out: int, approximate: str = "none", bias: bool = True):
+        super().__init__()
+        self.proj = nn.Linear(dim_in, dim_out, bias=bias)
+        self.approximate = approximate
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.proj(hidden_states)
+        hidden_states = F.gelu(hidden_states, approximate=self.approximate)
+        return hidden_states
+
+
+class FeedForward(nn.Module):
+    def __init__(
+        self,
+        dim: int,
+        dim_out: Optional[int] = None,
+        mult: int = 4,
+        dropout: float = 0.0,
+        activation_fn: str = "geglu",
+        final_dropout: bool = False,
+        inner_dim=None,
+        bias: bool = True,
+    ):
+        super().__init__()
+        if inner_dim is None:
+            inner_dim = int(dim * mult)
+        dim_out = dim_out if dim_out is not None else dim
+
+        # Build activation + projection matching diffusers pattern
+        if activation_fn == "gelu":
+            act_fn = GELU(dim, inner_dim, bias=bias)
+        elif activation_fn == "gelu-approximate":
+            act_fn = GELU(dim, inner_dim, approximate="tanh", bias=bias)
+        else:
+            act_fn = GELU(dim, inner_dim, bias=bias)
+
+        self.net = nn.ModuleList([])
+        self.net.append(act_fn)
+        self.net.append(nn.Dropout(dropout))
+        self.net.append(nn.Linear(inner_dim, dim_out, bias=bias))
+        if final_dropout:
+            self.net.append(nn.Dropout(dropout))
+
+    def forward(self, hidden_states: torch.Tensor, *args, **kwargs) -> torch.Tensor:
+        for module in self.net:
+            hidden_states = module(hidden_states)
+        return hidden_states
+
+
+def _to_tuple(x, dim=2):
+    if isinstance(x, int):
+        return (x,) * dim
+    elif len(x) == dim:
+        return x
+    else:
+        raise ValueError(f"Expected length {dim} or int, but got {x}")
+
+
+def get_meshgrid_nd(start, *args, dim=2):
+    if len(args) == 0:
+        num = _to_tuple(start, dim=dim)
+        start = (0,) * dim
+        stop = num
+    elif len(args) == 1:
+        start = _to_tuple(start, dim=dim)
+        stop = _to_tuple(args[0], dim=dim)
+        num = [stop[i] - start[i] for i in range(dim)]
+    elif len(args) == 2:
+        start = _to_tuple(start, dim=dim)
+        stop = _to_tuple(args[0], dim=dim)
+        num = _to_tuple(args[1], dim=dim)
+    else:
+        raise ValueError(f"len(args) should be 0, 1 or 2, but got {len(args)}")
+    axis_grid = []
+    for i in range(dim):
+        a, b, n = start[i], stop[i], num[i]
+        g = torch.linspace(a, b, n + 1, dtype=torch.float32)[:n]
+        axis_grid.append(g)
+    grid = torch.meshgrid(*axis_grid, indexing="ij")
+    grid = torch.stack(grid, dim=0)
+    return grid
+
+
+def reshape_for_broadcast(freqs_cis, x, head_first=False):
+    ndim = x.ndim
+    assert 0 <= 1 < ndim
+    if isinstance(freqs_cis, tuple):
+        if head_first:
+            assert freqs_cis[0].shape == (x.shape[-2], x.shape[-1])
+            shape = [d if i == ndim - 2 or i == ndim - 1 else 1 for i, d in enumerate(x.shape)]
+        else:
+            assert freqs_cis[0].shape == (x.shape[1], x.shape[-1])
+            shape = [d if i == 1 or i == ndim - 1 else 1 for i, d in enumerate(x.shape)]
+        return freqs_cis[0].view(*shape), freqs_cis[1].view(*shape)
+    else:
+        if head_first:
+            assert freqs_cis.shape == (x.shape[-2], x.shape[-1])
+            shape = [d if i == ndim - 2 or i == ndim - 1 else 1 for i, d in enumerate(x.shape)]
+        else:
+            assert freqs_cis.shape == (x.shape[1], x.shape[-1])
+            shape = [d if i == 1 or i == ndim - 1 else 1 for i, d in enumerate(x.shape)]
+        return freqs_cis.view(*shape)
+
+
+def rotate_half(x):
+    x_real, x_imag = x.float().reshape(*x.shape[:-1], -1, 2).unbind(-1)
+    return torch.stack([-x_imag, x_real], dim=-1).flatten(3)
+
+
+def apply_rotary_emb(xq, xk, freqs_cis, head_first=False):
+    cos, sin = reshape_for_broadcast(freqs_cis, xq, head_first)
+    cos, sin = cos.to(xq.device), sin.to(xq.device)
+    xq_out = (xq.float() * cos + rotate_half(xq.float()) * sin).type_as(xq)
+    xk_out = (xk.float() * cos + rotate_half(xk.float()) * sin).type_as(xk)
+    return xq_out, xk_out
+
+
+def get_1d_rotary_pos_embed(dim, pos, theta=10000.0, use_real=False, theta_rescale_factor=1.0, interpolation_factor=1.0):
+    if isinstance(pos, int):
+        pos = torch.arange(pos).float()
+    if theta_rescale_factor != 1.0:
+        theta *= theta_rescale_factor ** (dim / (dim - 2))
+    freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim))
+    freqs = torch.outer(pos * interpolation_factor, freqs)
+    if use_real:
+        freqs_cos = freqs.cos().repeat_interleave(2, dim=1)
+        freqs_sin = freqs.sin().repeat_interleave(2, dim=1)
+        return freqs_cos, freqs_sin
+    else:
+        return torch.polar(torch.ones_like(freqs), freqs)
+
+
+def get_nd_rotary_pos_embed(rope_dim_list, start, *args, theta=10000.0, use_real=False,
+                            txt_rope_size=None, theta_rescale_factor=1.0, interpolation_factor=1.0):
+    grid = get_meshgrid_nd(start, *args, dim=len(rope_dim_list))
+    if isinstance(theta_rescale_factor, (int, float)):
+        theta_rescale_factor = [theta_rescale_factor] * len(rope_dim_list)
+    elif isinstance(theta_rescale_factor, list) and len(theta_rescale_factor) == 1:
+        theta_rescale_factor = [theta_rescale_factor[0]] * len(rope_dim_list)
+    if isinstance(interpolation_factor, (int, float)):
+        interpolation_factor = [interpolation_factor] * len(rope_dim_list)
+    elif isinstance(interpolation_factor, list) and len(interpolation_factor) == 1:
+        interpolation_factor = [interpolation_factor[0]] * len(rope_dim_list)
+    embs = []
+    for i in range(len(rope_dim_list)):
+        emb = get_1d_rotary_pos_embed(
+            rope_dim_list[i], grid[i].reshape(-1), theta,
+            use_real=use_real, theta_rescale_factor=theta_rescale_factor[i],
+            interpolation_factor=interpolation_factor[i],
+        )
+        embs.append(emb)
+    if use_real:
+        vis_emb = (torch.cat([emb[0] for emb in embs], dim=1), torch.cat([emb[1] for emb in embs], dim=1))
+    else:
+        vis_emb = torch.cat(embs, dim=1)
+    if txt_rope_size is not None:
+        embs_txt = []
+        vis_max_ids = grid.view(-1).max().item()
+        grid_txt = torch.arange(txt_rope_size) + vis_max_ids + 1
+        for i in range(len(rope_dim_list)):
+            emb = get_1d_rotary_pos_embed(
+                rope_dim_list[i], grid_txt, theta,
+                use_real=use_real, theta_rescale_factor=theta_rescale_factor[i],
+                interpolation_factor=interpolation_factor[i],
+            )
+            embs_txt.append(emb)
+        if use_real:
+            txt_emb = (torch.cat([emb[0] for emb in embs_txt], dim=1), torch.cat([emb[1] for emb in embs_txt], dim=1))
+        else:
+            txt_emb = torch.cat(embs_txt, dim=1)
+    else:
+        txt_emb = None
+    return vis_emb, txt_emb
+
+
+class ModulateWan(nn.Module):
+    def __init__(self, hidden_size: int, factor: int, dtype=None, device=None):
+        super().__init__()
+        self.factor = factor
+        factory_kwargs = {"dtype": dtype, "device": device}
+        self.modulate_table = nn.Parameter(
+            torch.zeros(1, factor, hidden_size, **factory_kwargs) / hidden_size**0.5,
+            requires_grad=True
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        if len(x.shape) != 3:
+            x = x.unsqueeze(1)
+        return [o.squeeze(1) for o in (self.modulate_table + x).chunk(self.factor, dim=1)]
+
+
+def modulate(x, shift=None, scale=None):
+    if scale is None and shift is None:
+        return x
+    elif shift is None:
+        return x * (1 + scale.unsqueeze(1))
+    elif scale is None:
+        return x + shift.unsqueeze(1)
+    else:
+        return x * (1 + scale.unsqueeze(1)) + shift.unsqueeze(1)
+
+
+def apply_gate(x, gate=None, tanh=False):
+    if gate is None:
+        return x
+    if tanh:
+        return x * gate.unsqueeze(1).tanh()
+    else:
+        return x * gate.unsqueeze(1)
+
+
+def load_modulation(modulate_type: str, hidden_size: int, factor: int, dtype=None, device=None):
+    factory_kwargs = {"dtype": dtype, "device": device}
+    if modulate_type == 'wanx':
+        return ModulateWan(hidden_size, factor, **factory_kwargs)
+    raise ValueError(f"Unknown modulation type: {modulate_type}. Only 'wanx' is supported.")
+
+
+class RMSNorm(nn.Module):
+    def __init__(self, dim: int, elementwise_affine=True, eps: float = 1e-6, device=None, dtype=None):
+        factory_kwargs = {"device": device, "dtype": dtype}
+        super().__init__()
+        self.eps = eps
+        if elementwise_affine:
+            self.weight = nn.Parameter(torch.ones(dim, **factory_kwargs))
+
+    def _norm(self, x):
+        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)
+
+    def forward(self, x):
+        output = self._norm(x.float()).type_as(x)
+        if hasattr(self, "weight"):
+            output = output * self.weight
+        return output
+
+
+class MMDoubleStreamBlock(nn.Module):
+    """
+    A multimodal dit block with separate modulation for
+    text and image/video, see more details (SD3): https://arxiv.org/abs/2403.03206
+                                     (Flux.1): https://github.com/black-forest-labs/flux
+    """
+
+    def __init__(
+        self,
+        hidden_size: int,
+        heads_num: int,
+        mlp_width_ratio: float,
+        mlp_act_type: str = "gelu_tanh",
+        dtype: Optional[torch.dtype] = None,
+        device: Optional[torch.device] = None,
+        dit_modulation_type: Optional[str] = "wanx",
+    ):
+        factory_kwargs = {"device": device, "dtype": dtype}
+        super().__init__()
+        self.dit_modulation_type = dit_modulation_type
+        self.heads_num = heads_num
+        head_dim = hidden_size // heads_num
+        mlp_hidden_dim = int(hidden_size * mlp_width_ratio)
+
+        self.img_mod = load_modulation(
+            modulate_type=self.dit_modulation_type,
+            hidden_size=hidden_size, factor=6, **factory_kwargs,
+        )
+        self.img_norm1 = nn.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, **factory_kwargs)
+        self.img_attn_qkv = nn.Linear(hidden_size, hidden_size * 3, bias=True, **factory_kwargs)
+        self.img_attn_q_norm = RMSNorm(head_dim, elementwise_affine=True, eps=1e-6, **factory_kwargs)
+        self.img_attn_k_norm = RMSNorm(head_dim, elementwise_affine=True, eps=1e-6, **factory_kwargs)
+        self.img_attn_proj = nn.Linear(hidden_size, hidden_size, bias=True, **factory_kwargs)
+        self.img_norm2 = nn.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, **factory_kwargs)
+        self.img_mlp = FeedForward(hidden_size, inner_dim=mlp_hidden_dim, activation_fn="gelu-approximate")
+
+        self.txt_mod = load_modulation(
+            modulate_type=self.dit_modulation_type,
+            hidden_size=hidden_size, factor=6, **factory_kwargs,
+        )
+        self.txt_norm1 = nn.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, **factory_kwargs)
+        self.txt_attn_qkv = nn.Linear(hidden_size, hidden_size * 3, bias=True, **factory_kwargs)
+        self.txt_attn_q_norm = RMSNorm(head_dim, elementwise_affine=True, eps=1e-6, **factory_kwargs)
+        self.txt_attn_k_norm = RMSNorm(head_dim, elementwise_affine=True, eps=1e-6, **factory_kwargs)
+        self.txt_attn_proj = nn.Linear(hidden_size, hidden_size, bias=True, **factory_kwargs)
+        self.txt_norm2 = nn.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6, **factory_kwargs)
+        self.txt_mlp = FeedForward(hidden_size, inner_dim=mlp_hidden_dim, activation_fn="gelu-approximate")
+
+    def forward(
+        self,
+        img: torch.Tensor,
+        txt: torch.Tensor,
+        vec: torch.Tensor,
+        vis_freqs_cis: tuple = None,
+        txt_freqs_cis: tuple = None,
+        attn_kwargs: Optional[dict] = {},
+    ) -> Tuple[torch.Tensor, torch.Tensor]:
+        (
+            img_mod1_shift, img_mod1_scale, img_mod1_gate,
+            img_mod2_shift, img_mod2_scale, img_mod2_gate,
+        ) = self.img_mod(vec)
+        (
+            txt_mod1_shift, txt_mod1_scale, txt_mod1_gate,
+            txt_mod2_shift, txt_mod2_scale, txt_mod2_gate,
+        ) = self.txt_mod(vec)
+
+        img_modulated = self.img_norm1(img)
+        img_modulated = modulate(img_modulated, shift=img_mod1_shift, scale=img_mod1_scale)
+        img_qkv = self.img_attn_qkv(img_modulated)
+        img_q, img_k, img_v = rearrange(img_qkv, "B L (K H D) -> K B L H D", K=3, H=self.heads_num)
+        img_q = self.img_attn_q_norm(img_q).to(img_v)
+        img_k = self.img_attn_k_norm(img_k).to(img_v)
+
+        if vis_freqs_cis is not None:
+            img_qq, img_kk = apply_rotary_emb(img_q, img_k, vis_freqs_cis, head_first=False)
+            img_q, img_k = img_qq, img_kk
+
+        txt_modulated = self.txt_norm1(txt)
+        txt_modulated = modulate(txt_modulated, shift=txt_mod1_shift, scale=txt_mod1_scale)
+        txt_qkv = self.txt_attn_qkv(txt_modulated)
+        txt_q, txt_k, txt_v = rearrange(txt_qkv, "B L (K H D) -> K B L H D", K=3, H=self.heads_num)
+        txt_q = self.txt_attn_q_norm(txt_q).to(txt_v)
+        txt_k = self.txt_attn_k_norm(txt_k).to(txt_v)
+
+        if txt_freqs_cis is not None:
+            raise NotImplementedError("RoPE text is not supported for inference")
+
+        q = torch.cat((img_q, txt_q), dim=1)
+        k = torch.cat((img_k, txt_k), dim=1)
+        v = torch.cat((img_v, txt_v), dim=1)
+
+        # Use DiffSynth unified attention
+        attn_out = attention_forward(
+            q, k, v,
+            q_pattern="b s n d", k_pattern="b s n d", v_pattern="b s n d", out_pattern="b s n d",
+        )
+
+        attn_out = attn_out.flatten(2, 3)
+        img_attn, txt_attn = attn_out[:, : img.shape[1]], attn_out[:, img.shape[1]:]
+
+        img = img + apply_gate(self.img_attn_proj(img_attn), gate=img_mod1_gate)
+        img = img + apply_gate(
+            self.img_mlp(modulate(self.img_norm2(img), shift=img_mod2_shift, scale=img_mod2_scale)),
+            gate=img_mod2_gate,
+        )
+
+        txt = txt + apply_gate(self.txt_attn_proj(txt_attn), gate=txt_mod1_gate)
+        txt = txt + apply_gate(
+            self.txt_mlp(modulate(self.txt_norm2(txt), shift=txt_mod2_shift, scale=txt_mod2_scale)),
+            gate=txt_mod2_gate,
+        )
+
+        return img, txt
+
+
+class WanTimeTextImageEmbedding(nn.Module):
+    def __init__(
+        self,
+        dim: int,
+        time_freq_dim: int,
+        time_proj_dim: int,
+        text_embed_dim: int,
+        image_embed_dim: Optional[int] = None,
+        pos_embed_seq_len: Optional[int] = None,
+    ):
+        super().__init__()
+        self.timesteps_proj = Timesteps(num_channels=time_freq_dim, flip_sin_to_cos=True, downscale_freq_shift=0)
+        self.time_embedder = TimestepEmbedding(in_channels=time_freq_dim, time_embed_dim=dim)
+        self.act_fn = nn.SiLU()
+        self.time_proj = nn.Linear(dim, time_proj_dim)
+        self.text_embedder = PixArtAlphaTextProjection(text_embed_dim, dim, act_fn="gelu_tanh")
+
+    def forward(self, timestep: torch.Tensor, encoder_hidden_states: torch.Tensor):
+        timestep = self.timesteps_proj(timestep)
+        time_embedder_dtype = next(iter(self.time_embedder.parameters())).dtype
+        if timestep.dtype != time_embedder_dtype and time_embedder_dtype != torch.int8:
+            timestep = timestep.to(time_embedder_dtype)
+        temb = self.time_embedder(timestep).type_as(encoder_hidden_states)
+        timestep_proj = self.time_proj(self.act_fn(temb))
+        encoder_hidden_states = self.text_embedder(encoder_hidden_states)
+        return temb, timestep_proj, encoder_hidden_states
+
+
+class JoyAIImageDiT(nn.Module):
+    _supports_gradient_checkpointing = True
+
+    def __init__(
+        self,
+        patch_size: list = [1, 2, 2],
+        in_channels: int = 16,
+        out_channels: int = 16,
+        hidden_size: int = 4096,
+        heads_num: int = 32,
+        text_states_dim: int = 4096,
+        mlp_width_ratio: float = 4.0,
+        mm_double_blocks_depth: int = 40,
+        rope_dim_list: List[int] = [16, 56, 56],
+        rope_type: str = 'rope',
+        dtype: Optional[torch.dtype] = None,
+        device: Optional[torch.device] = None,
+        dit_modulation_type: str = "wanx",
+        theta: int = 10000,
+    ):
+        super().__init__()
+        self.out_channels = out_channels or in_channels
+        self.patch_size = patch_size
+        self.hidden_size = hidden_size
+        self.heads_num = heads_num
+        self.rope_dim_list = rope_dim_list
+        self.dit_modulation_type = dit_modulation_type
+        self.mm_double_blocks_depth = mm_double_blocks_depth
+        self.rope_type = rope_type
+        self.theta = theta
+
+        factory_kwargs = {"device": device, "dtype": dtype}
+
+        if hidden_size % heads_num != 0:
+            raise ValueError(f"Hidden size {hidden_size} must be divisible by heads_num {heads_num}")
+
+        self.img_in = nn.Conv3d(in_channels, hidden_size, kernel_size=patch_size, stride=patch_size)
+
+        self.condition_embedder = WanTimeTextImageEmbedding(
+            dim=hidden_size,
+            time_freq_dim=256,
+            time_proj_dim=hidden_size * 6,
+            text_embed_dim=text_states_dim,
+        )
+
+        self.double_blocks = nn.ModuleList([
+            MMDoubleStreamBlock(
+                self.hidden_size, self.heads_num,
+                mlp_width_ratio=mlp_width_ratio,
+                dit_modulation_type=self.dit_modulation_type,
+                **factory_kwargs,
+            )
+            for _ in range(mm_double_blocks_depth)
+        ])
+
+        self.norm_out = nn.LayerNorm(hidden_size, elementwise_affine=False, eps=1e-6)
+        self.proj_out = nn.Linear(hidden_size, self.out_channels * math.prod(patch_size), **factory_kwargs)
+
+    def get_rotary_pos_embed(self, vis_rope_size, txt_rope_size=None):
+        target_ndim = 3
+        if len(vis_rope_size) != target_ndim:
+            vis_rope_size = [1] * (target_ndim - len(vis_rope_size)) + vis_rope_size
+        head_dim = self.hidden_size // self.heads_num
+        rope_dim_list = self.rope_dim_list
+        if rope_dim_list is None:
+            rope_dim_list = [head_dim // target_ndim for _ in range(target_ndim)]
+        assert sum(rope_dim_list) == head_dim
+        vis_freqs, txt_freqs = get_nd_rotary_pos_embed(
+            rope_dim_list, vis_rope_size,
+            txt_rope_size=txt_rope_size if self.rope_type == 'mrope' else None,
+            theta=self.theta, use_real=True, theta_rescale_factor=1,
+        )
+        return vis_freqs, txt_freqs
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        timestep: torch.Tensor,
+        encoder_hidden_states: torch.Tensor = None,
+        encoder_hidden_states_mask: torch.Tensor = None,
+        return_dict: bool = True,
+        use_gradient_checkpointing: bool = False,
+        use_gradient_checkpointing_offload: bool = False,
+    ) -> Union[torch.Tensor, Dict[str, torch.Tensor]]:
+        is_multi_item = (len(hidden_states.shape) == 6)
+        num_items = 0
+        if is_multi_item:
+            num_items = hidden_states.shape[1]
+            if num_items > 1:
+                assert self.patch_size[0] == 1, "For multi-item input, patch_size[0] must be 1"
+                hidden_states = torch.cat([hidden_states[:, -1:], hidden_states[:, :-1]], dim=1)
+            hidden_states = rearrange(hidden_states, 'b n c t h w -> b c (n t) h w')
+
+        batch_size, _, ot, oh, ow = hidden_states.shape
+        tt, th, tw = ot // self.patch_size[0], oh // self.patch_size[1], ow // self.patch_size[2]
+
+        if encoder_hidden_states_mask is None:
+            encoder_hidden_states_mask = torch.ones(
+                (encoder_hidden_states.shape[0], encoder_hidden_states.shape[1]),
+                dtype=torch.bool,
+            ).to(encoder_hidden_states.device)
+
+        img = self.img_in(hidden_states).flatten(2).transpose(1, 2)
+        temb, vec, txt = self.condition_embedder(timestep, encoder_hidden_states)
+        if vec.shape[-1] > self.hidden_size:
+            vec = vec.unflatten(1, (6, -1))
+
+        txt_seq_len = txt.shape[1]
+        img_seq_len = img.shape[1]
+
+        vis_freqs_cis, txt_freqs_cis = self.get_rotary_pos_embed(
+            vis_rope_size=(tt, th, tw),
+            txt_rope_size=txt_seq_len if self.rope_type == 'mrope' else None,
+        )
+
+        for block in self.double_blocks:
+            img, txt = gradient_checkpoint_forward(
+                block,
+                use_gradient_checkpointing=use_gradient_checkpointing,
+                use_gradient_checkpointing_offload=use_gradient_checkpointing_offload,
+                img=img, txt=txt, vec=vec,
+                vis_freqs_cis=vis_freqs_cis, txt_freqs_cis=txt_freqs_cis,
+                attn_kwargs={},
+            )
+
+        img_len = img.shape[1]
+        x = torch.cat((img, txt), 1)
+        img = x[:, :img_len, ...]
+
+        img = self.proj_out(self.norm_out(img))
+        img = self.unpatchify(img, tt, th, tw)
+
+        if is_multi_item:
+            img = rearrange(img, 'b c (n t) h w -> b n c t h w', n=num_items)
+            if num_items > 1:
+                img = torch.cat([img[:, 1:], img[:, :1]], dim=1)
+
+        return img
+
+    def unpatchify(self, x, t, h, w):
+        c = self.out_channels
+        pt, ph, pw = self.patch_size
+        assert t * h * w == x.shape[1]
+        x = x.reshape(shape=(x.shape[0], t, h, w, pt, ph, pw, c))
+        x = torch.einsum("nthwopqc->nctohpwq", x)
+        return x.reshape(shape=(x.shape[0], c, t * pt, h * ph, w * pw))

diffsynth/models/joyai_image_text_encoder.py

@@ -0,0 +1,82 @@
+import torch
+from typing import Optional
+
+
+class JoyAIImageTextEncoder(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+        from transformers import Qwen3VLConfig, Qwen3VLForConditionalGeneration
+
+        config = Qwen3VLConfig(
+            text_config={
+                "attention_bias": False,
+                "attention_dropout": 0.0,
+                "bos_token_id": 151643,
+                "eos_token_id": 151645,
+                "head_dim": 128,
+                "hidden_act": "silu",
+                "hidden_size": 4096,
+                "initializer_range": 0.02,
+                "intermediate_size": 12288,
+                "max_position_embeddings": 262144,
+                "model_type": "qwen3_vl_text",
+                "num_attention_heads": 32,
+                "num_hidden_layers": 36,
+                "num_key_value_heads": 8,
+                "rms_norm_eps": 1e-6,
+                "rope_scaling": {
+                    "mrope_interleaved": True,
+                    "mrope_section": [24, 20, 20],
+                    "rope_type": "default",
+                },
+                "rope_theta": 5000000,
+                "use_cache": True,
+                "vocab_size": 151936,
+            },
+            vision_config={
+                "deepstack_visual_indexes": [8, 16, 24],
+                "depth": 27,
+                "hidden_act": "gelu_pytorch_tanh",
+                "hidden_size": 1152,
+                "in_channels": 3,
+                "initializer_range": 0.02,
+                "intermediate_size": 4304,
+                "model_type": "qwen3_vl",
+                "num_heads": 16,
+                "num_position_embeddings": 2304,
+                "out_hidden_size": 4096,
+                "patch_size": 16,
+                "spatial_merge_size": 2,
+                "temporal_patch_size": 2,
+            },
+            image_token_id=151655,
+            video_token_id=151656,
+            vision_start_token_id=151652,
+            vision_end_token_id=151653,
+            tie_word_embeddings=False,
+        )
+
+        self.model = Qwen3VLForConditionalGeneration(config)
+        self.config = config
+
+    def forward(
+        self,
+        input_ids: torch.LongTensor = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        pixel_values: Optional[torch.Tensor] = None,
+        image_grid_thw: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ):
+        pre_norm_output = [None]
+        def hook_fn(module, args, kwargs_output=None):
+            pre_norm_output[0] = args[0]
+        self.model.model.language_model.norm.register_forward_hook(hook_fn)
+        _ = self.model(
+            input_ids=input_ids,
+            pixel_values=pixel_values,
+            image_grid_thw=image_grid_thw,
+            attention_mask=attention_mask,
+            output_hidden_states=True,
+            **kwargs,
+        )
+        return pre_norm_output[0]

diffsynth/models/ltx2_dit.py

@@ -1280,6 +1280,7 @@ class LTXModel(torch.nn.Module):
     LTX model transformer implementation.
     This class implements the transformer blocks for the LTX model.
     """
+    _repeated_blocks = ["BasicAVTransformerBlock"]
 
     def __init__(  # noqa: PLR0913
         self,

diffsynth/models/qwen_image_dit.py

@@ -549,6 +549,9 @@ class QwenImageTransformerBlock(nn.Module):
 
 
 class QwenImageDiT(torch.nn.Module):
+
+    _repeated_blocks = ["QwenImageTransformerBlock"]
+
     def __init__(
         self,
         num_layers: int = 60,

diffsynth/models/stable_diffusion_text_encoder.py

@@ -0,0 +1,78 @@
+import torch
+
+
+class SDTextEncoder(torch.nn.Module):
+    def __init__(
+        self,
+        hidden_size=768,
+        intermediate_size=3072,
+        num_hidden_layers=12,
+        num_attention_heads=12,
+        max_position_embeddings=77,
+        vocab_size=49408,
+        layer_norm_eps=1e-05,
+        hidden_act="quick_gelu",
+        initializer_factor=1.0,
+        initializer_range=0.02,
+        bos_token_id=0,
+        eos_token_id=2,
+        pad_token_id=1,
+        projection_dim=768,
+    ):
+        super().__init__()
+        from transformers import CLIPConfig, CLIPTextModel
+
+        config = CLIPConfig(
+            text_config={
+                "hidden_size": hidden_size,
+                "intermediate_size": intermediate_size,
+                "num_hidden_layers": num_hidden_layers,
+                "num_attention_heads": num_attention_heads,
+                "max_position_embeddings": max_position_embeddings,
+                "vocab_size": vocab_size,
+                "layer_norm_eps": layer_norm_eps,
+                "hidden_act": hidden_act,
+                "initializer_factor": initializer_factor,
+                "initializer_range": initializer_range,
+                "bos_token_id": bos_token_id,
+                "eos_token_id": eos_token_id,
+                "pad_token_id": pad_token_id,
+                "projection_dim": projection_dim,
+                "dropout": 0.0,
+            },
+            vision_config={
+                "hidden_size": hidden_size,
+                "intermediate_size": intermediate_size,
+                "num_hidden_layers": num_hidden_layers,
+                "num_attention_heads": num_attention_heads,
+                "max_position_embeddings": max_position_embeddings,
+                "layer_norm_eps": layer_norm_eps,
+                "hidden_act": hidden_act,
+                "initializer_factor": initializer_factor,
+                "initializer_range": initializer_range,
+                "projection_dim": projection_dim,
+            },
+            projection_dim=projection_dim,
+        )
+        self.model = CLIPTextModel(config.text_config)
+        self.config = config
+
+    def forward(
+        self,
+        input_ids=None,
+        attention_mask=None,
+        position_ids=None,
+        output_hidden_states=True,
+        **kwargs,
+    ):
+        outputs = self.model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            output_hidden_states=output_hidden_states,
+            return_dict=True,
+            **kwargs,
+        )
+        if output_hidden_states:
+            return outputs.last_hidden_state, outputs.hidden_states
+        return outputs.last_hidden_state

diffsynth/models/stable_diffusion_unet.py

@@ -0,0 +1,912 @@
+# Copyright 2025 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import math
+from typing import Optional
+
+
+# ===== Time Embedding =====
+
+class Timesteps(nn.Module):
+    def __init__(self, num_channels, flip_sin_to_cos=True, freq_shift=0):
+        super().__init__()
+        self.num_channels = num_channels
+        self.flip_sin_to_cos = flip_sin_to_cos
+        self.freq_shift = freq_shift
+
+    def forward(self, timesteps):
+        half_dim = self.num_channels // 2
+        exponent = -math.log(10000) * torch.arange(half_dim, dtype=torch.float32, device=timesteps.device)
+        exponent = exponent / half_dim + self.freq_shift
+        emb = torch.exp(exponent)
+        emb = timesteps[:, None].float() * emb[None, :]
+        sin_emb = torch.sin(emb)
+        cos_emb = torch.cos(emb)
+        if self.flip_sin_to_cos:
+            emb = torch.cat([cos_emb, sin_emb], dim=-1)
+        else:
+            emb = torch.cat([sin_emb, cos_emb], dim=-1)
+        return emb
+
+
+class TimestepEmbedding(nn.Module):
+    def __init__(self, in_channels, time_embed_dim, act_fn="silu", out_dim=None):
+        super().__init__()
+        self.linear_1 = nn.Linear(in_channels, time_embed_dim)
+        self.act = nn.SiLU() if act_fn == "silu" else nn.GELU()
+        out_dim = out_dim if out_dim is not None else time_embed_dim
+        self.linear_2 = nn.Linear(time_embed_dim, out_dim)
+
+    def forward(self, sample):
+        sample = self.linear_1(sample)
+        sample = self.act(sample)
+        sample = self.linear_2(sample)
+        return sample
+
+
+# ===== ResNet Blocks =====
+
+class ResnetBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels=None,
+        conv_shortcut=False,
+        dropout=0.0,
+        temb_channels=512,
+        groups=32,
+        groups_out=None,
+        pre_norm=True,
+        eps=1e-6,
+        non_linearity="swish",
+        time_embedding_norm="default",
+        output_scale_factor=1.0,
+        use_in_shortcut=None,
+    ):
+        super().__init__()
+        self.pre_norm = pre_norm
+        self.time_embedding_norm = time_embedding_norm
+        self.output_scale_factor = output_scale_factor
+
+        if groups_out is None:
+            groups_out = groups
+
+        self.norm1 = nn.GroupNorm(num_groups=groups, num_channels=in_channels, eps=eps)
+        self.conv1 = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=3, stride=1, padding=1)
+
+        if temb_channels is not None:
+            if self.time_embedding_norm == "default":
+                self.time_emb_proj = nn.Linear(temb_channels, out_channels or in_channels)
+            elif self.time_embedding_norm == "scale_shift":
+                self.time_emb_proj = nn.Linear(temb_channels, 2 * (out_channels or in_channels))
+
+        self.norm2 = nn.GroupNorm(num_groups=groups_out, num_channels=out_channels or in_channels, eps=eps)
+        self.dropout = nn.Dropout(dropout)
+        self.conv2 = nn.Conv2d(out_channels or in_channels, out_channels or in_channels, kernel_size=3, stride=1, padding=1)
+
+        if non_linearity == "swish":
+            self.nonlinearity = nn.SiLU()
+        elif non_linearity == "silu":
+            self.nonlinearity = nn.SiLU()
+        elif non_linearity == "gelu":
+            self.nonlinearity = nn.GELU()
+        elif non_linearity == "relu":
+            self.nonlinearity = nn.ReLU()
+
+        self.use_conv_shortcut = conv_shortcut
+        self.conv_shortcut = None
+        if conv_shortcut:
+            self.conv_shortcut = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=1, stride=1, padding=0)
+        else:
+            self.conv_shortcut = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=1, stride=1, padding=0) if in_channels != (out_channels or in_channels) else None
+
+    def forward(self, input_tensor, temb=None):
+        hidden_states = input_tensor
+        hidden_states = self.norm1(hidden_states)
+        hidden_states = self.nonlinearity(hidden_states)
+        hidden_states = self.conv1(hidden_states)
+
+        if temb is not None:
+            temb = self.nonlinearity(temb)
+            temb = self.time_emb_proj(temb).unsqueeze(-1).unsqueeze(-1)
+
+        if temb is not None and self.time_embedding_norm == "default":
+            hidden_states = hidden_states + temb
+
+        hidden_states = self.norm2(hidden_states)
+
+        if temb is not None and self.time_embedding_norm == "scale_shift":
+            scale, shift = torch.chunk(temb, 2, dim=1)
+            hidden_states = hidden_states * (1 + scale) + shift
+
+        hidden_states = self.nonlinearity(hidden_states)
+        hidden_states = self.dropout(hidden_states)
+        hidden_states = self.conv2(hidden_states)
+
+        if self.conv_shortcut is not None:
+            input_tensor = self.conv_shortcut(input_tensor)
+
+        output_tensor = (input_tensor + hidden_states) / self.output_scale_factor
+        return output_tensor
+
+
+# ===== Transformer Blocks =====
+
+class GEGLU(nn.Module):
+    def __init__(self, dim_in, dim_out):
+        super().__init__()
+        self.proj = nn.Linear(dim_in, dim_out * 2)
+
+    def forward(self, hidden_states):
+        hidden_states, gate = self.proj(hidden_states).chunk(2, dim=-1)
+        return hidden_states * F.gelu(gate)
+
+
+class FeedForward(nn.Module):
+    def __init__(self, dim, dim_out=None, dropout=0.0):
+        super().__init__()
+        self.net = nn.ModuleList([
+            GEGLU(dim, dim * 4),
+            nn.Dropout(dropout),
+            nn.Linear(dim * 4, dim if dim_out is None else dim_out),
+        ])
+
+    def forward(self, hidden_states):
+        for module in self.net:
+            hidden_states = module(hidden_states)
+        return hidden_states
+
+
+class Attention(nn.Module):
+    """Attention block matching diffusers checkpoint key format.
+    Keys: to_q.weight, to_k.weight, to_v.weight, to_out.0.weight, to_out.0.bias
+    """
+    def __init__(
+        self,
+        query_dim,
+        heads=8,
+        dim_head=64,
+        dropout=0.0,
+        bias=False,
+        upcast_attention=False,
+        cross_attention_dim=None,
+    ):
+        super().__init__()
+        inner_dim = dim_head * heads
+        self.heads = heads
+        self.inner_dim = inner_dim
+        self.cross_attention_dim = cross_attention_dim if cross_attention_dim is not None else query_dim
+
+        self.to_q = nn.Linear(query_dim, inner_dim, bias=bias)
+        self.to_k = nn.Linear(self.cross_attention_dim, inner_dim, bias=bias)
+        self.to_v = nn.Linear(self.cross_attention_dim, inner_dim, bias=bias)
+        self.to_out = nn.ModuleList([
+            nn.Linear(inner_dim, query_dim, bias=True),
+            nn.Dropout(dropout),
+        ])
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None):
+        # Query
+        query = self.to_q(hidden_states)
+        batch_size, seq_len, _ = query.shape
+
+        # Key/Value
+        if encoder_hidden_states is None:
+            encoder_hidden_states = hidden_states
+        key = self.to_k(encoder_hidden_states)
+        value = self.to_v(encoder_hidden_states)
+
+        # Reshape for multi-head attention
+        head_dim = self.inner_dim // self.heads
+        query = query.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+        key = key.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+        value = value.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+
+        # Scaled dot-product attention
+        hidden_states = F.scaled_dot_product_attention(
+            query, key, value, attn_mask=None, dropout_p=0.0, is_causal=False
+        )
+
+        # Reshape back
+        hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, self.inner_dim)
+        hidden_states = hidden_states.to(query.dtype)
+
+        # Output projection
+        hidden_states = self.to_out[0](hidden_states)
+        hidden_states = self.to_out[1](hidden_states)
+
+        return hidden_states
+
+
+class BasicTransformerBlock(nn.Module):
+    def __init__(
+        self,
+        dim,
+        n_heads,
+        d_head,
+        dropout=0.0,
+        cross_attention_dim=None,
+        upcast_attention=False,
+    ):
+        super().__init__()
+        self.norm1 = nn.LayerNorm(dim)
+        self.attn1 = Attention(
+            query_dim=dim,
+            heads=n_heads,
+            dim_head=d_head,
+            dropout=dropout,
+            bias=False,
+            upcast_attention=upcast_attention,
+        )
+        self.norm2 = nn.LayerNorm(dim)
+        self.attn2 = Attention(
+            query_dim=dim,
+            heads=n_heads,
+            dim_head=d_head,
+            dropout=dropout,
+            bias=False,
+            upcast_attention=upcast_attention,
+            cross_attention_dim=cross_attention_dim,
+        )
+        self.norm3 = nn.LayerNorm(dim)
+        self.ff = FeedForward(dim, dropout=dropout)
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None):
+        # Self-attention
+        attn_output = self.attn1(self.norm1(hidden_states))
+        hidden_states = attn_output + hidden_states
+        # Cross-attention
+        attn_output = self.attn2(self.norm2(hidden_states), encoder_hidden_states=encoder_hidden_states)
+        hidden_states = attn_output + hidden_states
+        # Feed-forward
+        ff_output = self.ff(self.norm3(hidden_states))
+        hidden_states = ff_output + hidden_states
+        return hidden_states
+
+
+class Transformer2DModel(nn.Module):
+    """2D Transformer block wrapper matching diffusers checkpoint structure.
+    Keys: norm.weight/bias, proj_in.weight/bias, transformer_blocks.X.*, proj_out.weight/bias
+    """
+    def __init__(
+        self,
+        num_attention_heads=16,
+        attention_head_dim=64,
+        in_channels=320,
+        num_layers=1,
+        dropout=0.0,
+        norm_num_groups=32,
+        cross_attention_dim=768,
+        upcast_attention=False,
+    ):
+        super().__init__()
+        self.norm = nn.GroupNorm(num_groups=norm_num_groups, num_channels=in_channels, eps=1e-6)
+        self.proj_in = nn.Conv2d(in_channels, num_attention_heads * attention_head_dim, kernel_size=1, bias=True)
+
+        self.transformer_blocks = nn.ModuleList([
+            BasicTransformerBlock(
+                dim=num_attention_heads * attention_head_dim,
+                n_heads=num_attention_heads,
+                d_head=attention_head_dim,
+                dropout=dropout,
+                cross_attention_dim=cross_attention_dim,
+                upcast_attention=upcast_attention,
+            )
+            for _ in range(num_layers)
+        ])
+
+        self.proj_out = nn.Conv2d(num_attention_heads * attention_head_dim, in_channels, kernel_size=1, bias=True)
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None):
+        batch, channel, height, width = hidden_states.shape
+        residual = hidden_states
+
+        # Normalize and project to sequence
+        hidden_states = self.norm(hidden_states)
+        hidden_states = self.proj_in(hidden_states)
+        hidden_states = hidden_states.permute(0, 2, 3, 1).reshape(batch, -1, channel)
+
+        # Transformer blocks
+        for block in self.transformer_blocks:
+            hidden_states = block(hidden_states, encoder_hidden_states=encoder_hidden_states)
+
+        # Project back to 2D
+        hidden_states = hidden_states.reshape(batch, height, width, channel).permute(0, 3, 1, 2).contiguous()
+        hidden_states = self.proj_out(hidden_states)
+        hidden_states = hidden_states + residual
+        return hidden_states
+
+
+# ===== Down/Up Blocks =====
+
+class CrossAttnDownBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        transformer_layers_per_block=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        cross_attention_dim=768,
+        attention_head_dim=1,
+        downsample=True,
+    ):
+        super().__init__()
+        self.has_cross_attention = True
+
+        resnets = []
+        attentions = []
+
+        for i in range(num_layers):
+            in_channels_i = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels_i,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            attentions.append(
+                Transformer2DModel(
+                    num_attention_heads=attention_head_dim,
+                    attention_head_dim=out_channels // attention_head_dim,
+                    in_channels=out_channels,
+                    num_layers=transformer_layers_per_block,
+                    dropout=dropout,
+                    norm_num_groups=resnet_groups,
+                    cross_attention_dim=cross_attention_dim,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+        if downsample:
+            self.downsamplers = nn.ModuleList([
+                Downsample2D(out_channels, out_channels, padding=1)
+            ])
+        else:
+            self.downsamplers = None
+
+    def forward(self, hidden_states, temb=None, encoder_hidden_states=None):
+        output_states = []
+
+        for resnet, attn in zip(self.resnets, self.attentions):
+            hidden_states = resnet(hidden_states, temb)
+            hidden_states = attn(hidden_states, encoder_hidden_states=encoder_hidden_states)
+            output_states.append(hidden_states)
+
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states)
+            output_states.append(hidden_states)
+
+        return hidden_states, tuple(output_states)
+
+
+class DownBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        downsample=True,
+    ):
+        super().__init__()
+        self.has_cross_attention = False
+
+        resnets = []
+        for i in range(num_layers):
+            in_channels_i = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels_i,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+        self.resnets = nn.ModuleList(resnets)
+
+        if downsample:
+            self.downsamplers = nn.ModuleList([
+                Downsample2D(out_channels, out_channels, padding=1)
+            ])
+        else:
+            self.downsamplers = None
+
+    def forward(self, hidden_states, temb=None, encoder_hidden_states=None):
+        output_states = []
+        for resnet in self.resnets:
+            hidden_states = resnet(hidden_states, temb)
+            output_states.append(hidden_states)
+
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states)
+            output_states.append(hidden_states)
+
+        return hidden_states, tuple(output_states)
+
+
+class CrossAttnUpBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        prev_output_channel,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        transformer_layers_per_block=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        cross_attention_dim=768,
+        attention_head_dim=1,
+        upsample=True,
+    ):
+        super().__init__()
+        self.has_cross_attention = True
+
+        resnets = []
+        attentions = []
+
+        for i in range(num_layers):
+            res_skip_channels = in_channels if (i == num_layers - 1) else out_channels
+            resnet_in_channels = prev_output_channel if i == 0 else out_channels
+
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=resnet_in_channels + res_skip_channels,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            attentions.append(
+                Transformer2DModel(
+                    num_attention_heads=attention_head_dim,
+                    attention_head_dim=out_channels // attention_head_dim,
+                    in_channels=out_channels,
+                    num_layers=transformer_layers_per_block,
+                    dropout=dropout,
+                    norm_num_groups=resnet_groups,
+                    cross_attention_dim=cross_attention_dim,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+        if upsample:
+            self.upsamplers = nn.ModuleList([
+                Upsample2D(out_channels, out_channels)
+            ])
+        else:
+            self.upsamplers = None
+
+    def forward(self, hidden_states, res_hidden_states_tuple, temb=None, encoder_hidden_states=None, upsample_size=None):
+        for resnet, attn in zip(self.resnets, self.attentions):
+            # Pop res hidden states
+            res_hidden_states = res_hidden_states_tuple[-1]
+            res_hidden_states_tuple = res_hidden_states_tuple[:-1]
+            hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)
+            hidden_states = resnet(hidden_states, temb)
+            hidden_states = attn(hidden_states, encoder_hidden_states=encoder_hidden_states)
+
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states, upsample_size=upsample_size)
+
+        return hidden_states
+
+
+class UpBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        prev_output_channel,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        upsample=True,
+    ):
+        super().__init__()
+        self.has_cross_attention = False
+
+        resnets = []
+        for i in range(num_layers):
+            res_skip_channels = in_channels if (i == num_layers - 1) else out_channels
+            resnet_in_channels = prev_output_channel if i == 0 else out_channels
+
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=resnet_in_channels + res_skip_channels,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+        self.resnets = nn.ModuleList(resnets)
+
+        if upsample:
+            self.upsamplers = nn.ModuleList([
+                Upsample2D(out_channels, out_channels)
+            ])
+        else:
+            self.upsamplers = None
+
+    def forward(self, hidden_states, res_hidden_states_tuple, temb=None, encoder_hidden_states=None, upsample_size=None):
+        for resnet in self.resnets:
+            res_hidden_states = res_hidden_states_tuple[-1]
+            res_hidden_states_tuple = res_hidden_states_tuple[:-1]
+            hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)
+            hidden_states = resnet(hidden_states, temb)
+
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states, upsample_size=upsample_size)
+
+        return hidden_states
+
+
+# ===== UNet Mid Block =====
+
+class UNetMidBlock2DCrossAttn(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        transformer_layers_per_block=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        cross_attention_dim=768,
+        attention_head_dim=1,
+    ):
+        super().__init__()
+        resnet_groups = resnet_groups if resnet_groups is not None else min(in_channels // 4, 32)
+
+        # There is always at least one resnet
+        resnets = [
+            ResnetBlock2D(
+                in_channels=in_channels,
+                out_channels=in_channels,
+                temb_channels=temb_channels,
+                eps=resnet_eps,
+                groups=resnet_groups,
+                dropout=dropout,
+                time_embedding_norm=resnet_time_scale_shift,
+                non_linearity=resnet_act_fn,
+                output_scale_factor=1.0,
+                pre_norm=resnet_pre_norm,
+            )
+        ]
+        attentions = []
+
+        for _ in range(num_layers):
+            attentions.append(
+                Transformer2DModel(
+                    num_attention_heads=attention_head_dim,
+                    attention_head_dim=in_channels // attention_head_dim,
+                    in_channels=in_channels,
+                    num_layers=transformer_layers_per_block,
+                    dropout=dropout,
+                    norm_num_groups=resnet_groups,
+                    cross_attention_dim=cross_attention_dim,
+                )
+            )
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels,
+                    out_channels=in_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+    def forward(self, hidden_states, temb=None, encoder_hidden_states=None):
+        hidden_states = self.resnets[0](hidden_states, temb)
+        for attn, resnet in zip(self.attentions, self.resnets[1:]):
+            hidden_states = attn(hidden_states, encoder_hidden_states=encoder_hidden_states)
+            hidden_states = resnet(hidden_states, temb)
+        return hidden_states
+
+
+# ===== Downsample / Upsample =====
+
+class Downsample2D(nn.Module):
+    def __init__(self, in_channels, out_channels, padding=1):
+        super().__init__()
+        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=2, padding=padding)
+        self.padding = padding
+
+    def forward(self, hidden_states):
+        if self.padding == 0:
+            hidden_states = F.pad(hidden_states, (0, 1, 0, 1), mode="constant", value=0)
+        return self.conv(hidden_states)
+
+
+class Upsample2D(nn.Module):
+    def __init__(self, in_channels, out_channels):
+        super().__init__()
+        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1)
+
+    def forward(self, hidden_states, upsample_size=None):
+        if upsample_size is not None:
+            hidden_states = F.interpolate(hidden_states, size=upsample_size, mode="nearest")
+        else:
+            hidden_states = F.interpolate(hidden_states, scale_factor=2.0, mode="nearest")
+        return self.conv(hidden_states)
+
+
+# ===== UNet2DConditionModel =====
+
+class UNet2DConditionModel(nn.Module):
+    """Stable Diffusion UNet with cross-attention conditioning.
+    state_dict keys match the diffusers UNet2DConditionModel checkpoint format.
+    """
+    def __init__(
+        self,
+        sample_size=64,
+        in_channels=4,
+        out_channels=4,
+        down_block_types=("CrossAttnDownBlock2D", "CrossAttnDownBlock2D", "CrossAttnDownBlock2D", "DownBlock2D"),
+        up_block_types=("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D"),
+        block_out_channels=(320, 640, 1280, 1280),
+        layers_per_block=2,
+        cross_attention_dim=768,
+        attention_head_dim=8,
+        norm_num_groups=32,
+        norm_eps=1e-5,
+        dropout=0.0,
+        act_fn="silu",
+        time_embedding_type="positional",
+        flip_sin_to_cos=True,
+        freq_shift=0,
+        time_embedding_dim=None,
+        resnet_time_scale_shift="default",
+        upcast_attention=False,
+    ):
+        super().__init__()
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.sample_size = sample_size
+
+        # Time embedding
+        timestep_embedding_dim = time_embedding_dim or block_out_channels[0]
+        self.time_proj = Timesteps(timestep_embedding_dim, flip_sin_to_cos=flip_sin_to_cos, freq_shift=freq_shift)
+        time_embed_dim = block_out_channels[0] * 4
+        self.time_embedding = TimestepEmbedding(timestep_embedding_dim, time_embed_dim)
+
+        # Input
+        self.conv_in = nn.Conv2d(in_channels, block_out_channels[0], kernel_size=3, padding=1)
+
+        # Down blocks
+        self.down_blocks = nn.ModuleList()
+        output_channel = block_out_channels[0]
+        for i, down_block_type in enumerate(down_block_types):
+            input_channel = output_channel
+            output_channel = block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+
+            if "CrossAttn" in down_block_type:
+                down_block = CrossAttnDownBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block,
+                    transformer_layers_per_block=1,
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    attention_head_dim=attention_head_dim,
+                    downsample=not is_final_block,
+                )
+            else:
+                down_block = DownBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block,
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    downsample=not is_final_block,
+                )
+            self.down_blocks.append(down_block)
+
+        # Mid block
+        self.mid_block = UNetMidBlock2DCrossAttn(
+            in_channels=block_out_channels[-1],
+            temb_channels=time_embed_dim,
+            dropout=dropout,
+            num_layers=1,
+            transformer_layers_per_block=1,
+            resnet_eps=norm_eps,
+            resnet_time_scale_shift=resnet_time_scale_shift,
+            resnet_act_fn=act_fn,
+            resnet_groups=norm_num_groups,
+            cross_attention_dim=cross_attention_dim,
+            attention_head_dim=attention_head_dim,
+        )
+
+        # Up blocks
+        self.up_blocks = nn.ModuleList()
+        reversed_block_out_channels = list(reversed(block_out_channels))
+        output_channel = reversed_block_out_channels[0]
+
+        for i, up_block_type in enumerate(up_block_types):
+            prev_output_channel = output_channel
+            output_channel = reversed_block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+
+            # in_channels for up blocks: diffusers uses reversed_block_out_channels[min(i+1, len-1)]
+            input_channel = reversed_block_out_channels[min(i + 1, len(block_out_channels) - 1)]
+
+            if "CrossAttn" in up_block_type:
+                up_block = CrossAttnUpBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    prev_output_channel=prev_output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block + 1,
+                    transformer_layers_per_block=1,
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    attention_head_dim=attention_head_dim,
+                    upsample=not is_final_block,
+                )
+            else:
+                up_block = UpBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    prev_output_channel=prev_output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block + 1,
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    upsample=not is_final_block,
+                )
+            self.up_blocks.append(up_block)
+
+        # Output
+        self.conv_norm_out = nn.GroupNorm(num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=norm_eps)
+        self.conv_act = nn.SiLU()
+        self.conv_out = nn.Conv2d(block_out_channels[0], out_channels, kernel_size=3, padding=1)
+
+    def forward(self, sample, timestep, encoder_hidden_states, cross_attention_kwargs=None, timestep_cond=None, added_cond_kwargs=None, return_dict=True):
+        # 1. Time embedding
+        timesteps = timestep
+        if not torch.is_tensor(timesteps):
+            timesteps = torch.tensor([timesteps], dtype=torch.long, device=sample.device)
+        elif torch.is_tensor(timesteps) and len(timesteps.shape) == 0:
+            timesteps = timesteps[None].to(sample.device)
+
+        t_emb = self.time_proj(timesteps)
+        t_emb = t_emb.to(dtype=sample.dtype)
+        emb = self.time_embedding(t_emb)
+
+        # 2. Pre-process
+        sample = self.conv_in(sample)
+
+        # 3. Down
+        down_block_res_samples = (sample,)
+        for down_block in self.down_blocks:
+            sample, res_samples = down_block(
+                hidden_states=sample,
+                temb=emb,
+                encoder_hidden_states=encoder_hidden_states,
+            )
+            down_block_res_samples += res_samples
+
+        # 4. Mid
+        sample = self.mid_block(sample, emb, encoder_hidden_states=encoder_hidden_states)
+
+        # 5. Up
+        for up_block in self.up_blocks:
+            res_samples = down_block_res_samples[-len(up_block.resnets):]
+            down_block_res_samples = down_block_res_samples[:-len(up_block.resnets)]
+
+            upsample_size = down_block_res_samples[-1].shape[2:] if down_block_res_samples else None
+            sample = up_block(
+                hidden_states=sample,
+                temb=emb,
+                encoder_hidden_states=encoder_hidden_states,
+                res_hidden_states_tuple=res_samples,
+                upsample_size=upsample_size,
+            )
+
+        # 6. Post-process
+        sample = self.conv_norm_out(sample)
+        sample = self.conv_act(sample)
+        sample = self.conv_out(sample)
+
+        if not return_dict:
+            return (sample,)
+        return sample

diffsynth/models/stable_diffusion_vae.py

@@ -0,0 +1,642 @@
+# Copyright 2025 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import torch
+import torch.nn as nn
+from typing import Optional
+
+
+class DiagonalGaussianDistribution:
+    def __init__(self, parameters: torch.Tensor, deterministic: bool = False):
+        self.parameters = parameters
+        self.mean, self.logvar = torch.chunk(parameters, 2, dim=1)
+        self.logvar = torch.clamp(self.logvar, -30.0, 20.0)
+        self.deterministic = deterministic
+        self.std = torch.exp(0.5 * self.logvar)
+        self.var = torch.exp(self.logvar)
+        if self.deterministic:
+            self.var = self.std = torch.zeros_like(
+                self.mean, device=self.parameters.device, dtype=self.parameters.dtype
+            )
+
+    def sample(self, generator: Optional[torch.Generator] = None) -> torch.Tensor:
+        # randn_like doesn't accept generator on all torch versions
+        sample = torch.randn(self.mean.shape, generator=generator,
+                           device=self.parameters.device, dtype=self.parameters.dtype)
+        return self.mean + self.std * sample
+
+    def kl(self, other: Optional["DiagonalGaussianDistribution"] = None) -> torch.Tensor:
+        if self.deterministic:
+            return torch.tensor([0.0])
+        if other is None:
+            return 0.5 * torch.sum(
+                torch.pow(self.mean, 2) + self.var - 1.0 - self.logvar,
+                dim=[1, 2, 3],
+            )
+        return 0.5 * torch.sum(
+            torch.pow(self.mean - other.mean, 2) / other.var
+            + self.var / other.var - 1.0 - self.logvar + other.logvar,
+            dim=[1, 2, 3],
+        )
+
+    def mode(self) -> torch.Tensor:
+        return self.mean
+
+
+class ResnetBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels=None,
+        conv_shortcut=False,
+        dropout=0.0,
+        temb_channels=512,
+        groups=32,
+        groups_out=None,
+        pre_norm=True,
+        eps=1e-6,
+        non_linearity="swish",
+        time_embedding_norm="default",
+        output_scale_factor=1.0,
+        use_in_shortcut=None,
+    ):
+        super().__init__()
+        self.pre_norm = pre_norm
+        self.time_embedding_norm = time_embedding_norm
+        self.output_scale_factor = output_scale_factor
+
+        if groups_out is None:
+            groups_out = groups
+
+        self.norm1 = nn.GroupNorm(num_groups=groups, num_channels=in_channels, eps=eps)
+        self.conv1 = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=3, stride=1, padding=1)
+
+        if temb_channels is not None:
+            if self.time_embedding_norm == "default":
+                self.time_emb_proj = nn.Linear(temb_channels, out_channels or in_channels)
+            elif self.time_embedding_norm == "scale_shift":
+                self.time_emb_proj = nn.Linear(temb_channels, 2 * (out_channels or in_channels))
+
+        self.norm2 = nn.GroupNorm(num_groups=groups_out, num_channels=out_channels or in_channels, eps=eps)
+        self.dropout = nn.Dropout(dropout)
+        self.conv2 = nn.Conv2d(out_channels or in_channels, out_channels or in_channels, kernel_size=3, stride=1, padding=1)
+
+        if non_linearity == "swish":
+            self.nonlinearity = nn.SiLU()
+        elif non_linearity == "silu":
+            self.nonlinearity = nn.SiLU()
+        elif non_linearity == "gelu":
+            self.nonlinearity = nn.GELU()
+        elif non_linearity == "relu":
+            self.nonlinearity = nn.ReLU()
+        else:
+            raise ValueError(f"Unsupported non_linearity: {non_linearity}")
+
+        self.use_conv_shortcut = conv_shortcut
+        self.conv_shortcut = None
+        if conv_shortcut:
+            self.conv_shortcut = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=1, stride=1, padding=0)
+        else:
+            self.conv_shortcut = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=1, stride=1, padding=0) if in_channels != (out_channels or in_channels) else None
+
+    def forward(self, input_tensor, temb=None):
+        hidden_states = input_tensor
+        hidden_states = self.norm1(hidden_states)
+        hidden_states = self.nonlinearity(hidden_states)
+
+        hidden_states = self.conv1(hidden_states)
+
+        if temb is not None:
+            temb = self.nonlinearity(temb)
+            temb = self.time_emb_proj(temb).unsqueeze(-1).unsqueeze(-1)
+
+        if temb is not None and self.time_embedding_norm == "default":
+            hidden_states = hidden_states + temb
+
+        hidden_states = self.norm2(hidden_states)
+
+        if temb is not None and self.time_embedding_norm == "scale_shift":
+            scale, shift = torch.chunk(temb, 2, dim=1)
+            hidden_states = hidden_states * (1 + scale) + shift
+
+        hidden_states = self.nonlinearity(hidden_states)
+        hidden_states = self.dropout(hidden_states)
+        hidden_states = self.conv2(hidden_states)
+
+        if self.conv_shortcut is not None:
+            input_tensor = self.conv_shortcut(input_tensor)
+
+        output_tensor = (input_tensor + hidden_states) / self.output_scale_factor
+        return output_tensor
+
+
+class DownEncoderBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        output_scale_factor=1.0,
+        add_downsample=True,
+        downsample_padding=1,
+    ):
+        super().__init__()
+        resnets = []
+        for i in range(num_layers):
+            in_channels_i = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels_i,
+                    out_channels=out_channels,
+                    temb_channels=None,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=output_scale_factor,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+        self.resnets = nn.ModuleList(resnets)
+
+        if add_downsample:
+            self.downsamplers = nn.ModuleList([
+                Downsample2D(out_channels, out_channels, padding=downsample_padding)
+            ])
+        else:
+            self.downsamplers = None
+
+    def forward(self, hidden_states, *args, **kwargs):
+        for resnet in self.resnets:
+            hidden_states = resnet(hidden_states, temb=None)
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states)
+        return hidden_states
+
+
+class UpDecoderBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        output_scale_factor=1.0,
+        add_upsample=True,
+        temb_channels=None,
+    ):
+        super().__init__()
+        resnets = []
+        for i in range(num_layers):
+            in_channels_i = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels_i,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=output_scale_factor,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+        self.resnets = nn.ModuleList(resnets)
+
+        if add_upsample:
+            self.upsamplers = nn.ModuleList([
+                Upsample2D(out_channels, out_channels)
+            ])
+        else:
+            self.upsamplers = None
+
+    def forward(self, hidden_states, temb=None):
+        for resnet in self.resnets:
+            hidden_states = resnet(hidden_states, temb=temb)
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states)
+        return hidden_states
+
+
+class UNetMidBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        temb_channels=None,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        add_attention=True,
+        attention_head_dim=1,
+        output_scale_factor=1.0,
+    ):
+        super().__init__()
+        resnet_groups = resnet_groups if resnet_groups is not None else min(in_channels // 4, 32)
+        self.add_attention = add_attention
+
+        # there is always at least one resnet
+        resnets = [
+            ResnetBlock2D(
+                in_channels=in_channels,
+                out_channels=in_channels,
+                temb_channels=temb_channels,
+                eps=resnet_eps,
+                groups=resnet_groups,
+                dropout=dropout,
+                time_embedding_norm=resnet_time_scale_shift,
+                non_linearity=resnet_act_fn,
+                output_scale_factor=output_scale_factor,
+                pre_norm=resnet_pre_norm,
+            )
+        ]
+        attentions = []
+
+        if attention_head_dim is None:
+            attention_head_dim = in_channels
+
+        for _ in range(num_layers):
+            if self.add_attention:
+                attentions.append(
+                    AttentionBlock(
+                        in_channels,
+                        num_groups=resnet_groups,
+                        eps=resnet_eps,
+                    )
+                )
+            else:
+                attentions.append(None)
+
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels,
+                    out_channels=in_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=output_scale_factor,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+    def forward(self, hidden_states, temb=None):
+        hidden_states = self.resnets[0](hidden_states, temb)
+        for attn, resnet in zip(self.attentions, self.resnets[1:]):
+            if attn is not None:
+                hidden_states = attn(hidden_states)
+            hidden_states = resnet(hidden_states, temb)
+        return hidden_states
+
+
+class AttentionBlock(nn.Module):
+    """Simple attention block for VAE mid block.
+    Mirrors diffusers Attention class with AttnProcessor2_0 for VAE use case.
+    Uses modern key names (to_q, to_k, to_v, to_out) matching in-memory diffusers structure.
+    Checkpoint uses deprecated keys (query, key, value, proj_attn) — mapped via converter.
+    """
+    def __init__(self, channels, num_groups=32, eps=1e-6):
+        super().__init__()
+        self.channels = channels
+        self.eps = eps
+        self.heads = 1
+        self.rescale_output_factor = 1.0
+
+        self.group_norm = nn.GroupNorm(num_groups=num_groups, num_channels=channels, eps=eps, affine=True)
+        self.to_q = nn.Linear(channels, channels, bias=True)
+        self.to_k = nn.Linear(channels, channels, bias=True)
+        self.to_v = nn.Linear(channels, channels, bias=True)
+        self.to_out = nn.ModuleList([
+            nn.Linear(channels, channels, bias=True),
+            nn.Dropout(0.0),
+        ])
+
+    def forward(self, hidden_states):
+        residual = hidden_states
+
+        # Group norm
+        hidden_states = self.group_norm(hidden_states)
+
+        # Flatten spatial dims: (B, C, H, W) -> (B, H*W, C)
+        batch_size, channel, height, width = hidden_states.shape
+        hidden_states = hidden_states.view(batch_size, channel, height * width).transpose(1, 2)
+
+        # QKV projection
+        query = self.to_q(hidden_states)
+        key = self.to_k(hidden_states)
+        value = self.to_v(hidden_states)
+
+        # Reshape for attention: (B, seq, dim) -> (B, heads, seq, head_dim)
+        inner_dim = key.shape[-1]
+        head_dim = inner_dim // self.heads
+        query = query.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+        key = key.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+        value = value.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+
+        # Scaled dot-product attention
+        hidden_states = torch.nn.functional.scaled_dot_product_attention(
+            query, key, value, attn_mask=None, dropout_p=0.0, is_causal=False
+        )
+
+        # Reshape back: (B, heads, seq, head_dim) -> (B, seq, heads*head_dim)
+        hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, self.heads * head_dim)
+        hidden_states = hidden_states.to(query.dtype)
+
+        # Output projection + dropout
+        hidden_states = self.to_out[0](hidden_states)
+        hidden_states = self.to_out[1](hidden_states)
+
+        # Reshape back to 4D and add residual
+        hidden_states = hidden_states.transpose(-1, -2).reshape(batch_size, channel, height, width)
+        hidden_states = hidden_states + residual
+
+        # Rescale output factor
+        hidden_states = hidden_states / self.rescale_output_factor
+
+        return hidden_states
+
+
+class Downsample2D(nn.Module):
+    """Downsampling layer matching diffusers Downsample2D with use_conv=True.
+    Key names: conv.weight/bias.
+    When padding=0, applies explicit F.pad before conv to match dimension.
+    """
+    def __init__(self, in_channels, out_channels, padding=1):
+        super().__init__()
+        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=2, padding=0)
+        self.padding = padding
+
+    def forward(self, hidden_states):
+        if self.padding == 0:
+            import torch.nn.functional as F
+            hidden_states = F.pad(hidden_states, (0, 1, 0, 1), mode="constant", value=0)
+        return self.conv(hidden_states)
+
+
+class Upsample2D(nn.Module):
+    """Upsampling layer with key names matching diffusers checkpoint: conv.weight/bias."""
+    def __init__(self, in_channels, out_channels):
+        super().__init__()
+        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1)
+
+    def forward(self, hidden_states):
+        hidden_states = torch.nn.functional.interpolate(hidden_states, scale_factor=2.0, mode="nearest")
+        return self.conv(hidden_states)
+
+
+class Encoder(nn.Module):
+    def __init__(
+        self,
+        in_channels=3,
+        out_channels=3,
+        down_block_types=("DownEncoderBlock2D",),
+        block_out_channels=(64,),
+        layers_per_block=2,
+        norm_num_groups=32,
+        act_fn="silu",
+        double_z=True,
+        mid_block_add_attention=True,
+    ):
+        super().__init__()
+        self.layers_per_block = layers_per_block
+
+        self.conv_in = nn.Conv2d(in_channels, block_out_channels[0], kernel_size=3, stride=1, padding=1)
+
+        self.down_blocks = nn.ModuleList([])
+        output_channel = block_out_channels[0]
+        for i, down_block_type in enumerate(down_block_types):
+            input_channel = output_channel
+            output_channel = block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+            down_block = DownEncoderBlock2D(
+                in_channels=input_channel,
+                out_channels=output_channel,
+                num_layers=self.layers_per_block,
+                resnet_eps=1e-6,
+                resnet_act_fn=act_fn,
+                resnet_groups=norm_num_groups,
+                add_downsample=not is_final_block,
+                downsample_padding=0,
+            )
+            self.down_blocks.append(down_block)
+
+        # mid
+        self.mid_block = UNetMidBlock2D(
+            in_channels=block_out_channels[-1],
+            resnet_eps=1e-6,
+            resnet_act_fn=act_fn,
+            output_scale_factor=1,
+            resnet_time_scale_shift="default",
+            attention_head_dim=block_out_channels[-1],
+            resnet_groups=norm_num_groups,
+            temb_channels=None,
+            add_attention=mid_block_add_attention,
+        )
+
+        # out
+        self.conv_norm_out = nn.GroupNorm(num_channels=block_out_channels[-1], num_groups=norm_num_groups, eps=1e-6)
+        self.conv_act = nn.SiLU()
+        conv_out_channels = 2 * out_channels if double_z else out_channels
+        self.conv_out = nn.Conv2d(block_out_channels[-1], conv_out_channels, 3, padding=1)
+
+    def forward(self, sample):
+        sample = self.conv_in(sample)
+        for down_block in self.down_blocks:
+            sample = down_block(sample)
+        sample = self.mid_block(sample)
+        sample = self.conv_norm_out(sample)
+        sample = self.conv_act(sample)
+        sample = self.conv_out(sample)
+        return sample
+
+
+class Decoder(nn.Module):
+    def __init__(
+        self,
+        in_channels=3,
+        out_channels=3,
+        up_block_types=("UpDecoderBlock2D",),
+        block_out_channels=(64,),
+        layers_per_block=2,
+        norm_num_groups=32,
+        act_fn="silu",
+        norm_type="group",
+        mid_block_add_attention=True,
+    ):
+        super().__init__()
+        self.layers_per_block = layers_per_block
+
+        self.conv_in = nn.Conv2d(in_channels, block_out_channels[-1], kernel_size=3, stride=1, padding=1)
+
+        self.up_blocks = nn.ModuleList([])
+        temb_channels = in_channels if norm_type == "spatial" else None
+
+        # mid
+        self.mid_block = UNetMidBlock2D(
+            in_channels=block_out_channels[-1],
+            resnet_eps=1e-6,
+            resnet_act_fn=act_fn,
+            output_scale_factor=1,
+            resnet_time_scale_shift="default" if norm_type == "group" else norm_type,
+            attention_head_dim=block_out_channels[-1],
+            resnet_groups=norm_num_groups,
+            temb_channels=temb_channels,
+            add_attention=mid_block_add_attention,
+        )
+
+        # up
+        reversed_block_out_channels = list(reversed(block_out_channels))
+        output_channel = reversed_block_out_channels[0]
+        for i, up_block_type in enumerate(up_block_types):
+            prev_output_channel = output_channel
+            output_channel = reversed_block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+            up_block = UpDecoderBlock2D(
+                in_channels=prev_output_channel,
+                out_channels=output_channel,
+                num_layers=self.layers_per_block + 1,
+                resnet_eps=1e-6,
+                resnet_act_fn=act_fn,
+                resnet_groups=norm_num_groups,
+                add_upsample=not is_final_block,
+                temb_channels=temb_channels,
+            )
+            self.up_blocks.append(up_block)
+            prev_output_channel = output_channel
+
+        # out
+        self.conv_norm_out = nn.GroupNorm(num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=1e-6)
+        self.conv_act = nn.SiLU()
+        self.conv_out = nn.Conv2d(block_out_channels[0], out_channels, 3, padding=1)
+
+    def forward(self, sample, latent_embeds=None):
+        sample = self.conv_in(sample)
+        sample = self.mid_block(sample, latent_embeds)
+        for up_block in self.up_blocks:
+            sample = up_block(sample, latent_embeds)
+        sample = self.conv_norm_out(sample)
+        sample = self.conv_act(sample)
+        sample = self.conv_out(sample)
+        return sample
+
+
+class StableDiffusionVAE(nn.Module):
+    def __init__(
+        self,
+        in_channels=3,
+        out_channels=3,
+        down_block_types=("DownEncoderBlock2D", "DownEncoderBlock2D", "DownEncoderBlock2D", "DownEncoderBlock2D"),
+        up_block_types=("UpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D"),
+        block_out_channels=(128, 256, 512, 512),
+        layers_per_block=2,
+        act_fn="silu",
+        latent_channels=4,
+        norm_num_groups=32,
+        sample_size=512,
+        scaling_factor=0.18215,
+        shift_factor=None,
+        latents_mean=None,
+        latents_std=None,
+        force_upcast=True,
+        use_quant_conv=True,
+        use_post_quant_conv=True,
+        mid_block_add_attention=True,
+    ):
+        super().__init__()
+        self.encoder = Encoder(
+            in_channels=in_channels,
+            out_channels=latent_channels,
+            down_block_types=down_block_types,
+            block_out_channels=block_out_channels,
+            layers_per_block=layers_per_block,
+            norm_num_groups=norm_num_groups,
+            act_fn=act_fn,
+            double_z=True,
+            mid_block_add_attention=mid_block_add_attention,
+        )
+        self.decoder = Decoder(
+            in_channels=latent_channels,
+            out_channels=out_channels,
+            up_block_types=up_block_types,
+            block_out_channels=block_out_channels,
+            layers_per_block=layers_per_block,
+            norm_num_groups=norm_num_groups,
+            act_fn=act_fn,
+            mid_block_add_attention=mid_block_add_attention,
+        )
+
+        self.quant_conv = nn.Conv2d(2 * latent_channels, 2 * latent_channels, 1) if use_quant_conv else None
+        self.post_quant_conv = nn.Conv2d(latent_channels, latent_channels, 1) if use_post_quant_conv else None
+
+        self.latents_mean = latents_mean
+        self.latents_std = latents_std
+        self.scaling_factor = scaling_factor
+        self.shift_factor = shift_factor
+        self.sample_size = sample_size
+        self.force_upcast = force_upcast
+
+    def _encode(self, x):
+        h = self.encoder(x)
+        if self.quant_conv is not None:
+            h = self.quant_conv(h)
+        return h
+
+    def encode(self, x):
+        h = self._encode(x)
+        posterior = DiagonalGaussianDistribution(h)
+        return posterior
+
+    def _decode(self, z):
+        if self.post_quant_conv is not None:
+            z = self.post_quant_conv(z)
+        return self.decoder(z)
+
+    def decode(self, z):
+        return self._decode(z)
+
+    def forward(self, sample, sample_posterior=True, return_dict=True, generator=None):
+        posterior = self.encode(sample)
+        if sample_posterior:
+            z = posterior.sample(generator=generator)
+        else:
+            z = posterior.mode()
+        # Scale latent
+        z = z * self.scaling_factor
+        decode = self.decode(z)
+        if return_dict:
+            return {"sample": decode, "posterior": posterior, "latent_sample": z}
+        return decode, posterior

diffsynth/models/stable_diffusion_xl_text_encoder.py

@@ -0,0 +1,62 @@
+import torch
+
+
+class SDXLTextEncoder2(torch.nn.Module):
+    def __init__(
+        self,
+        hidden_size=1280,
+        intermediate_size=5120,
+        num_hidden_layers=32,
+        num_attention_heads=20,
+        max_position_embeddings=77,
+        vocab_size=49408,
+        layer_norm_eps=1e-05,
+        hidden_act="gelu",
+        initializer_factor=1.0,
+        initializer_range=0.02,
+        bos_token_id=0,
+        eos_token_id=2,
+        pad_token_id=1,
+        projection_dim=1280,
+    ):
+        super().__init__()
+        from transformers import CLIPTextConfig, CLIPTextModelWithProjection
+
+        config = CLIPTextConfig(
+            hidden_size=hidden_size,
+            intermediate_size=intermediate_size,
+            num_hidden_layers=num_hidden_layers,
+            num_attention_heads=num_attention_heads,
+            max_position_embeddings=max_position_embeddings,
+            vocab_size=vocab_size,
+            layer_norm_eps=layer_norm_eps,
+            hidden_act=hidden_act,
+            initializer_factor=initializer_factor,
+            initializer_range=initializer_range,
+            bos_token_id=bos_token_id,
+            eos_token_id=eos_token_id,
+            pad_token_id=pad_token_id,
+            projection_dim=projection_dim,
+        )
+        self.model = CLIPTextModelWithProjection(config)
+        self.config = config
+
+    def forward(
+        self,
+        input_ids=None,
+        attention_mask=None,
+        position_ids=None,
+        output_hidden_states=True,
+        **kwargs,
+    ):
+        outputs = self.model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            output_hidden_states=output_hidden_states,
+            return_dict=True,
+            **kwargs,
+        )
+        if output_hidden_states:
+            return outputs.text_embeds, outputs.hidden_states
+        return outputs.text_embeds

diffsynth/models/stable_diffusion_xl_unet.py

@@ -0,0 +1,922 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import math
+from typing import Optional
+
+
+# ===== Time Embedding =====
+
+class Timesteps(nn.Module):
+    def __init__(self, num_channels, flip_sin_to_cos=True, freq_shift=0):
+        super().__init__()
+        self.num_channels = num_channels
+        self.flip_sin_to_cos = flip_sin_to_cos
+        self.freq_shift = freq_shift
+
+    def forward(self, timesteps):
+        half_dim = self.num_channels // 2
+        exponent = -math.log(10000) * torch.arange(half_dim, dtype=torch.float32, device=timesteps.device)
+        exponent = exponent / half_dim + self.freq_shift
+        emb = torch.exp(exponent)
+        emb = timesteps[:, None].float() * emb[None, :]
+        sin_emb = torch.sin(emb)
+        cos_emb = torch.cos(emb)
+        if self.flip_sin_to_cos:
+            emb = torch.cat([cos_emb, sin_emb], dim=-1)
+        else:
+            emb = torch.cat([sin_emb, cos_emb], dim=-1)
+        return emb
+
+
+class TimestepEmbedding(nn.Module):
+    def __init__(self, in_channels, time_embed_dim, act_fn="silu", out_dim=None):
+        super().__init__()
+        self.linear_1 = nn.Linear(in_channels, time_embed_dim)
+        self.act = nn.SiLU() if act_fn == "silu" else nn.GELU()
+        out_dim = out_dim if out_dim is not None else time_embed_dim
+        self.linear_2 = nn.Linear(time_embed_dim, out_dim)
+
+    def forward(self, sample):
+        sample = self.linear_1(sample)
+        sample = self.act(sample)
+        sample = self.linear_2(sample)
+        return sample
+
+
+# ===== ResNet Blocks =====
+
+class ResnetBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels=None,
+        conv_shortcut=False,
+        dropout=0.0,
+        temb_channels=512,
+        groups=32,
+        groups_out=None,
+        pre_norm=True,
+        eps=1e-6,
+        non_linearity="swish",
+        time_embedding_norm="default",
+        output_scale_factor=1.0,
+        use_in_shortcut=None,
+    ):
+        super().__init__()
+        self.pre_norm = pre_norm
+        self.time_embedding_norm = time_embedding_norm
+        self.output_scale_factor = output_scale_factor
+
+        if groups_out is None:
+            groups_out = groups
+
+        self.norm1 = nn.GroupNorm(num_groups=groups, num_channels=in_channels, eps=eps)
+        self.conv1 = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=3, stride=1, padding=1)
+
+        if temb_channels is not None:
+            if self.time_embedding_norm == "default":
+                self.time_emb_proj = nn.Linear(temb_channels, out_channels or in_channels)
+            elif self.time_embedding_norm == "scale_shift":
+                self.time_emb_proj = nn.Linear(temb_channels, 2 * (out_channels or in_channels))
+
+        self.norm2 = nn.GroupNorm(num_groups=groups_out, num_channels=out_channels or in_channels, eps=eps)
+        self.dropout = nn.Dropout(dropout)
+        self.conv2 = nn.Conv2d(out_channels or in_channels, out_channels or in_channels, kernel_size=3, stride=1, padding=1)
+
+        if non_linearity == "swish":
+            self.nonlinearity = nn.SiLU()
+        elif non_linearity == "silu":
+            self.nonlinearity = nn.SiLU()
+        elif non_linearity == "gelu":
+            self.nonlinearity = nn.GELU()
+        elif non_linearity == "relu":
+            self.nonlinearity = nn.ReLU()
+
+        self.use_conv_shortcut = conv_shortcut
+        self.conv_shortcut = None
+        if conv_shortcut:
+            self.conv_shortcut = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=1, stride=1, padding=0)
+        else:
+            self.conv_shortcut = nn.Conv2d(in_channels, out_channels or in_channels, kernel_size=1, stride=1, padding=0) if in_channels != (out_channels or in_channels) else None
+
+    def forward(self, input_tensor, temb=None):
+        hidden_states = input_tensor
+        hidden_states = self.norm1(hidden_states)
+        hidden_states = self.nonlinearity(hidden_states)
+        hidden_states = self.conv1(hidden_states)
+
+        if temb is not None:
+            temb = self.nonlinearity(temb)
+            temb = self.time_emb_proj(temb).unsqueeze(-1).unsqueeze(-1)
+
+        if temb is not None and self.time_embedding_norm == "default":
+            hidden_states = hidden_states + temb
+
+        hidden_states = self.norm2(hidden_states)
+
+        if temb is not None and self.time_embedding_norm == "scale_shift":
+            scale, shift = torch.chunk(temb, 2, dim=1)
+            hidden_states = hidden_states * (1 + scale) + shift
+
+        hidden_states = self.nonlinearity(hidden_states)
+        hidden_states = self.dropout(hidden_states)
+        hidden_states = self.conv2(hidden_states)
+
+        if self.conv_shortcut is not None:
+            input_tensor = self.conv_shortcut(input_tensor)
+
+        output_tensor = (input_tensor + hidden_states) / self.output_scale_factor
+        return output_tensor
+
+
+# ===== Transformer Blocks =====
+
+class GEGLU(nn.Module):
+    def __init__(self, dim_in, dim_out):
+        super().__init__()
+        self.proj = nn.Linear(dim_in, dim_out * 2)
+
+    def forward(self, hidden_states):
+        hidden_states, gate = self.proj(hidden_states).chunk(2, dim=-1)
+        return hidden_states * F.gelu(gate)
+
+
+class FeedForward(nn.Module):
+    def __init__(self, dim, dim_out=None, dropout=0.0):
+        super().__init__()
+        self.net = nn.ModuleList([
+            GEGLU(dim, dim * 4),
+            nn.Dropout(dropout),
+            nn.Linear(dim * 4, dim if dim_out is None else dim_out),
+        ])
+
+    def forward(self, hidden_states):
+        for module in self.net:
+            hidden_states = module(hidden_states)
+        return hidden_states
+
+
+class Attention(nn.Module):
+    def __init__(
+        self,
+        query_dim,
+        heads=8,
+        dim_head=64,
+        dropout=0.0,
+        bias=False,
+        upcast_attention=False,
+        cross_attention_dim=None,
+    ):
+        super().__init__()
+        inner_dim = dim_head * heads
+        self.heads = heads
+        self.inner_dim = inner_dim
+        self.cross_attention_dim = cross_attention_dim if cross_attention_dim is not None else query_dim
+
+        self.to_q = nn.Linear(query_dim, inner_dim, bias=bias)
+        self.to_k = nn.Linear(self.cross_attention_dim, inner_dim, bias=bias)
+        self.to_v = nn.Linear(self.cross_attention_dim, inner_dim, bias=bias)
+        self.to_out = nn.ModuleList([
+            nn.Linear(inner_dim, query_dim, bias=True),
+            nn.Dropout(dropout),
+        ])
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None):
+        query = self.to_q(hidden_states)
+        batch_size, seq_len, _ = query.shape
+
+        if encoder_hidden_states is None:
+            encoder_hidden_states = hidden_states
+        key = self.to_k(encoder_hidden_states)
+        value = self.to_v(encoder_hidden_states)
+
+        head_dim = self.inner_dim // self.heads
+        query = query.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+        key = key.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+        value = value.view(batch_size, -1, self.heads, head_dim).transpose(1, 2)
+
+        hidden_states = F.scaled_dot_product_attention(
+            query, key, value, attn_mask=None, dropout_p=0.0, is_causal=False
+        )
+
+        hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, self.inner_dim)
+        hidden_states = hidden_states.to(query.dtype)
+
+        hidden_states = self.to_out[0](hidden_states)
+        hidden_states = self.to_out[1](hidden_states)
+
+        return hidden_states
+
+
+class BasicTransformerBlock(nn.Module):
+    def __init__(
+        self,
+        dim,
+        n_heads,
+        d_head,
+        dropout=0.0,
+        cross_attention_dim=None,
+        upcast_attention=False,
+    ):
+        super().__init__()
+        self.norm1 = nn.LayerNorm(dim)
+        self.attn1 = Attention(
+            query_dim=dim,
+            heads=n_heads,
+            dim_head=d_head,
+            dropout=dropout,
+            bias=False,
+            upcast_attention=upcast_attention,
+        )
+        self.norm2 = nn.LayerNorm(dim)
+        self.attn2 = Attention(
+            query_dim=dim,
+            heads=n_heads,
+            dim_head=d_head,
+            dropout=dropout,
+            bias=False,
+            upcast_attention=upcast_attention,
+            cross_attention_dim=cross_attention_dim,
+        )
+        self.norm3 = nn.LayerNorm(dim)
+        self.ff = FeedForward(dim, dropout=dropout)
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None):
+        attn_output = self.attn1(self.norm1(hidden_states))
+        hidden_states = attn_output + hidden_states
+        attn_output = self.attn2(self.norm2(hidden_states), encoder_hidden_states=encoder_hidden_states)
+        hidden_states = attn_output + hidden_states
+        ff_output = self.ff(self.norm3(hidden_states))
+        hidden_states = ff_output + hidden_states
+        return hidden_states
+
+
+class Transformer2DModel(nn.Module):
+    def __init__(
+        self,
+        num_attention_heads=16,
+        attention_head_dim=64,
+        in_channels=320,
+        num_layers=1,
+        dropout=0.0,
+        norm_num_groups=32,
+        cross_attention_dim=768,
+        upcast_attention=False,
+        use_linear_projection=False,
+    ):
+        super().__init__()
+        self.num_attention_heads = num_attention_heads
+        self.attention_head_dim = attention_head_dim
+        inner_dim = num_attention_heads * attention_head_dim
+        self.use_linear_projection = use_linear_projection
+
+        self.norm = nn.GroupNorm(num_groups=norm_num_groups, num_channels=in_channels, eps=1e-6)
+
+        if use_linear_projection:
+            self.proj_in = nn.Linear(in_channels, inner_dim, bias=True)
+        else:
+            self.proj_in = nn.Conv2d(in_channels, inner_dim, kernel_size=1, bias=True)
+
+        self.transformer_blocks = nn.ModuleList([
+            BasicTransformerBlock(
+                dim=inner_dim,
+                n_heads=num_attention_heads,
+                d_head=attention_head_dim,
+                dropout=dropout,
+                cross_attention_dim=cross_attention_dim,
+                upcast_attention=upcast_attention,
+            )
+            for _ in range(num_layers)
+        ])
+
+        if use_linear_projection:
+            self.proj_out = nn.Linear(inner_dim, in_channels, bias=True)
+        else:
+            self.proj_out = nn.Conv2d(inner_dim, in_channels, kernel_size=1, bias=True)
+
+    def forward(self, hidden_states, encoder_hidden_states=None, attention_mask=None):
+        batch, channel, height, width = hidden_states.shape
+        residual = hidden_states
+
+        hidden_states = self.norm(hidden_states)
+
+        if self.use_linear_projection:
+            hidden_states = hidden_states.permute(0, 2, 3, 1).reshape(batch, -1, channel)
+            hidden_states = self.proj_in(hidden_states)
+        else:
+            hidden_states = self.proj_in(hidden_states)
+            hidden_states = hidden_states.permute(0, 2, 3, 1).reshape(batch, -1, channel)
+
+        for block in self.transformer_blocks:
+            hidden_states = block(hidden_states, encoder_hidden_states=encoder_hidden_states)
+
+        if self.use_linear_projection:
+            hidden_states = self.proj_out(hidden_states)
+            hidden_states = hidden_states.reshape(batch, height, width, channel).permute(0, 3, 1, 2).contiguous()
+        else:
+            hidden_states = hidden_states.reshape(batch, height, width, channel).permute(0, 3, 1, 2).contiguous()
+            hidden_states = self.proj_out(hidden_states)
+
+        hidden_states = hidden_states + residual
+        return hidden_states
+
+
+# ===== Down/Up Blocks =====
+
+class CrossAttnDownBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        transformer_layers_per_block=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        cross_attention_dim=768,
+        attention_head_dim=1,
+        downsample=True,
+        use_linear_projection=False,
+    ):
+        super().__init__()
+        self.has_cross_attention = True
+
+        resnets = []
+        attentions = []
+
+        for i in range(num_layers):
+            in_channels_i = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels_i,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            attentions.append(
+                Transformer2DModel(
+                    num_attention_heads=attention_head_dim,
+                    attention_head_dim=out_channels // attention_head_dim,
+                    in_channels=out_channels,
+                    num_layers=transformer_layers_per_block,
+                    dropout=dropout,
+                    norm_num_groups=resnet_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    use_linear_projection=use_linear_projection,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+        if downsample:
+            self.downsamplers = nn.ModuleList([
+                Downsample2D(out_channels, out_channels, padding=1)
+            ])
+        else:
+            self.downsamplers = None
+
+    def forward(self, hidden_states, temb=None, encoder_hidden_states=None):
+        output_states = []
+
+        for resnet, attn in zip(self.resnets, self.attentions):
+            hidden_states = resnet(hidden_states, temb)
+            hidden_states = attn(hidden_states, encoder_hidden_states=encoder_hidden_states)
+            output_states.append(hidden_states)
+
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states)
+            output_states.append(hidden_states)
+
+        return hidden_states, tuple(output_states)
+
+
+class DownBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        downsample=True,
+    ):
+        super().__init__()
+        self.has_cross_attention = False
+
+        resnets = []
+        for i in range(num_layers):
+            in_channels_i = in_channels if i == 0 else out_channels
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels_i,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+        self.resnets = nn.ModuleList(resnets)
+
+        if downsample:
+            self.downsamplers = nn.ModuleList([
+                Downsample2D(out_channels, out_channels, padding=1)
+            ])
+        else:
+            self.downsamplers = None
+
+    def forward(self, hidden_states, temb=None, encoder_hidden_states=None):
+        output_states = []
+        for resnet in self.resnets:
+            hidden_states = resnet(hidden_states, temb)
+            output_states.append(hidden_states)
+
+        if self.downsamplers is not None:
+            for downsampler in self.downsamplers:
+                hidden_states = downsampler(hidden_states)
+            output_states.append(hidden_states)
+
+        return hidden_states, tuple(output_states)
+
+
+class CrossAttnUpBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        prev_output_channel,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        transformer_layers_per_block=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        cross_attention_dim=768,
+        attention_head_dim=1,
+        upsample=True,
+        use_linear_projection=False,
+    ):
+        super().__init__()
+        self.has_cross_attention = True
+
+        resnets = []
+        attentions = []
+
+        for i in range(num_layers):
+            res_skip_channels = in_channels if (i == num_layers - 1) else out_channels
+            resnet_in_channels = prev_output_channel if i == 0 else out_channels
+
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=resnet_in_channels + res_skip_channels,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+            attentions.append(
+                Transformer2DModel(
+                    num_attention_heads=attention_head_dim,
+                    attention_head_dim=out_channels // attention_head_dim,
+                    in_channels=out_channels,
+                    num_layers=transformer_layers_per_block,
+                    dropout=dropout,
+                    norm_num_groups=resnet_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    use_linear_projection=use_linear_projection,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+        if upsample:
+            self.upsamplers = nn.ModuleList([
+                Upsample2D(out_channels, out_channels)
+            ])
+        else:
+            self.upsamplers = None
+
+    def forward(self, hidden_states, res_hidden_states_tuple, temb=None, encoder_hidden_states=None, upsample_size=None):
+        for resnet, attn in zip(self.resnets, self.attentions):
+            res_hidden_states = res_hidden_states_tuple[-1]
+            res_hidden_states_tuple = res_hidden_states_tuple[:-1]
+            hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)
+            hidden_states = resnet(hidden_states, temb)
+            hidden_states = attn(hidden_states, encoder_hidden_states=encoder_hidden_states)
+
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states, upsample_size=upsample_size)
+
+        return hidden_states
+
+
+class UpBlock2D(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        out_channels,
+        prev_output_channel,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        upsample=True,
+    ):
+        super().__init__()
+        self.has_cross_attention = False
+
+        resnets = []
+        for i in range(num_layers):
+            res_skip_channels = in_channels if (i == num_layers - 1) else out_channels
+            resnet_in_channels = prev_output_channel if i == 0 else out_channels
+
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=resnet_in_channels + res_skip_channels,
+                    out_channels=out_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+        self.resnets = nn.ModuleList(resnets)
+
+        if upsample:
+            self.upsamplers = nn.ModuleList([
+                Upsample2D(out_channels, out_channels)
+            ])
+        else:
+            self.upsamplers = None
+
+    def forward(self, hidden_states, res_hidden_states_tuple, temb=None, encoder_hidden_states=None, upsample_size=None):
+        for resnet in self.resnets:
+            res_hidden_states = res_hidden_states_tuple[-1]
+            res_hidden_states_tuple = res_hidden_states_tuple[:-1]
+            hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)
+            hidden_states = resnet(hidden_states, temb)
+
+        if self.upsamplers is not None:
+            for upsampler in self.upsamplers:
+                hidden_states = upsampler(hidden_states, upsample_size=upsample_size)
+
+        return hidden_states
+
+
+# ===== UNet Mid Block =====
+
+class UNetMidBlock2DCrossAttn(nn.Module):
+    def __init__(
+        self,
+        in_channels,
+        temb_channels=1280,
+        dropout=0.0,
+        num_layers=1,
+        transformer_layers_per_block=1,
+        resnet_eps=1e-6,
+        resnet_time_scale_shift="default",
+        resnet_act_fn="swish",
+        resnet_groups=32,
+        resnet_pre_norm=True,
+        cross_attention_dim=768,
+        attention_head_dim=1,
+        use_linear_projection=False,
+    ):
+        super().__init__()
+        resnet_groups = resnet_groups if resnet_groups is not None else min(in_channels // 4, 32)
+
+        resnets = [
+            ResnetBlock2D(
+                in_channels=in_channels,
+                out_channels=in_channels,
+                temb_channels=temb_channels,
+                eps=resnet_eps,
+                groups=resnet_groups,
+                dropout=dropout,
+                time_embedding_norm=resnet_time_scale_shift,
+                non_linearity=resnet_act_fn,
+                output_scale_factor=1.0,
+                pre_norm=resnet_pre_norm,
+            )
+        ]
+        attentions = []
+
+        for _ in range(num_layers):
+            attentions.append(
+                Transformer2DModel(
+                    num_attention_heads=attention_head_dim,
+                    attention_head_dim=in_channels // attention_head_dim,
+                    in_channels=in_channels,
+                    num_layers=transformer_layers_per_block,
+                    dropout=dropout,
+                    norm_num_groups=resnet_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    use_linear_projection=use_linear_projection,
+                )
+            )
+            resnets.append(
+                ResnetBlock2D(
+                    in_channels=in_channels,
+                    out_channels=in_channels,
+                    temb_channels=temb_channels,
+                    eps=resnet_eps,
+                    groups=resnet_groups,
+                    dropout=dropout,
+                    time_embedding_norm=resnet_time_scale_shift,
+                    non_linearity=resnet_act_fn,
+                    output_scale_factor=1.0,
+                    pre_norm=resnet_pre_norm,
+                )
+            )
+
+        self.attentions = nn.ModuleList(attentions)
+        self.resnets = nn.ModuleList(resnets)
+
+    def forward(self, hidden_states, temb=None, encoder_hidden_states=None):
+        hidden_states = self.resnets[0](hidden_states, temb)
+        for attn, resnet in zip(self.attentions, self.resnets[1:]):
+            hidden_states = attn(hidden_states, encoder_hidden_states=encoder_hidden_states)
+            hidden_states = resnet(hidden_states, temb)
+        return hidden_states
+
+
+# ===== Downsample / Upsample =====
+
+class Downsample2D(nn.Module):
+    def __init__(self, in_channels, out_channels, padding=1):
+        super().__init__()
+        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=2, padding=padding)
+        self.padding = padding
+
+    def forward(self, hidden_states):
+        if self.padding == 0:
+            hidden_states = F.pad(hidden_states, (0, 1, 0, 1), mode="constant", value=0)
+        return self.conv(hidden_states)
+
+
+class Upsample2D(nn.Module):
+    def __init__(self, in_channels, out_channels):
+        super().__init__()
+        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1)
+
+    def forward(self, hidden_states, upsample_size=None):
+        if upsample_size is not None:
+            hidden_states = F.interpolate(hidden_states, size=upsample_size, mode="nearest")
+        else:
+            hidden_states = F.interpolate(hidden_states, scale_factor=2.0, mode="nearest")
+        return self.conv(hidden_states)
+
+
+# ===== SDXL UNet2DConditionModel =====
+
+class SDXLUNet2DConditionModel(nn.Module):
+    def __init__(
+        self,
+        sample_size=128,
+        in_channels=4,
+        out_channels=4,
+        down_block_types=("DownBlock2D", "CrossAttnDownBlock2D", "CrossAttnDownBlock2D"),
+        up_block_types=("CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "UpBlock2D"),
+        block_out_channels=(320, 640, 1280),
+        layers_per_block=2,
+        cross_attention_dim=2048,
+        attention_head_dim=5,
+        transformer_layers_per_block=1,
+        norm_num_groups=32,
+        norm_eps=1e-5,
+        dropout=0.0,
+        act_fn="silu",
+        time_embedding_type="positional",
+        flip_sin_to_cos=True,
+        freq_shift=0,
+        time_embedding_dim=None,
+        resnet_time_scale_shift="default",
+        upcast_attention=False,
+        use_linear_projection=False,
+        addition_embed_type=None,
+        addition_time_embed_dim=None,
+        projection_class_embeddings_input_dim=None,
+    ):
+        super().__init__()
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.sample_size = sample_size
+        self.addition_embed_type = addition_embed_type
+
+        if isinstance(attention_head_dim, int):
+            attention_head_dim = (attention_head_dim,) * len(down_block_types)
+        if isinstance(transformer_layers_per_block, int):
+            transformer_layers_per_block = [transformer_layers_per_block] * len(down_block_types)
+
+        timestep_embedding_dim = time_embedding_dim or block_out_channels[0]
+        self.time_proj = Timesteps(timestep_embedding_dim, flip_sin_to_cos=flip_sin_to_cos, freq_shift=freq_shift)
+        time_embed_dim = block_out_channels[0] * 4
+        self.time_embedding = TimestepEmbedding(timestep_embedding_dim, time_embed_dim)
+
+        if addition_embed_type == "text_time":
+            self.add_time_proj = Timesteps(addition_time_embed_dim, flip_sin_to_cos=flip_sin_to_cos, freq_shift=freq_shift)
+            self.add_embedding = TimestepEmbedding(projection_class_embeddings_input_dim, time_embed_dim)
+
+        self.conv_in = nn.Conv2d(in_channels, block_out_channels[0], kernel_size=3, padding=1)
+
+        self.down_blocks = nn.ModuleList()
+        output_channel = block_out_channels[0]
+        for i, down_block_type in enumerate(down_block_types):
+            input_channel = output_channel
+            output_channel = block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+
+            if "CrossAttn" in down_block_type:
+                down_block = CrossAttnDownBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block,
+                    transformer_layers_per_block=transformer_layers_per_block[i],
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    attention_head_dim=attention_head_dim[i],
+                    downsample=not is_final_block,
+                    use_linear_projection=use_linear_projection,
+                )
+            else:
+                down_block = DownBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block,
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    downsample=not is_final_block,
+                )
+            self.down_blocks.append(down_block)
+
+        self.mid_block = UNetMidBlock2DCrossAttn(
+            in_channels=block_out_channels[-1],
+            temb_channels=time_embed_dim,
+            dropout=dropout,
+            num_layers=1,
+            transformer_layers_per_block=transformer_layers_per_block[-1],
+            resnet_eps=norm_eps,
+            resnet_time_scale_shift=resnet_time_scale_shift,
+            resnet_act_fn=act_fn,
+            resnet_groups=norm_num_groups,
+            cross_attention_dim=cross_attention_dim,
+            attention_head_dim=attention_head_dim[-1],
+            use_linear_projection=use_linear_projection,
+        )
+
+        self.up_blocks = nn.ModuleList()
+        reversed_block_out_channels = list(reversed(block_out_channels))
+        reversed_attention_head_dim = list(reversed(attention_head_dim))
+        reversed_transformer_layers_per_block = list(reversed(transformer_layers_per_block))
+        output_channel = reversed_block_out_channels[0]
+
+        for i, up_block_type in enumerate(up_block_types):
+            prev_output_channel = output_channel
+            output_channel = reversed_block_out_channels[i]
+            is_final_block = i == len(block_out_channels) - 1
+
+            input_channel = reversed_block_out_channels[min(i + 1, len(block_out_channels) - 1)]
+
+            if "CrossAttn" in up_block_type:
+                up_block = CrossAttnUpBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    prev_output_channel=prev_output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block + 1,
+                    transformer_layers_per_block=reversed_transformer_layers_per_block[i],
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    cross_attention_dim=cross_attention_dim,
+                    attention_head_dim=reversed_attention_head_dim[i],
+                    upsample=not is_final_block,
+                    use_linear_projection=use_linear_projection,
+                )
+            else:
+                up_block = UpBlock2D(
+                    in_channels=input_channel,
+                    out_channels=output_channel,
+                    prev_output_channel=prev_output_channel,
+                    temb_channels=time_embed_dim,
+                    dropout=dropout,
+                    num_layers=layers_per_block + 1,
+                    resnet_eps=norm_eps,
+                    resnet_time_scale_shift=resnet_time_scale_shift,
+                    resnet_act_fn=act_fn,
+                    resnet_groups=norm_num_groups,
+                    upsample=not is_final_block,
+                )
+            self.up_blocks.append(up_block)
+
+        self.conv_norm_out = nn.GroupNorm(num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=norm_eps)
+        self.conv_act = nn.SiLU()
+        self.conv_out = nn.Conv2d(block_out_channels[0], out_channels, kernel_size=3, padding=1)
+
+    def forward(self, sample, timestep, encoder_hidden_states, cross_attention_kwargs=None, timestep_cond=None, added_cond_kwargs=None, return_dict=True):
+        timesteps = timestep
+        if not torch.is_tensor(timesteps):
+            timesteps = torch.tensor([timesteps], dtype=torch.long, device=sample.device)
+        elif torch.is_tensor(timesteps) and len(timesteps.shape) == 0:
+            timesteps = timesteps[None].to(sample.device)
+
+        t_emb = self.time_proj(timesteps)
+        t_emb = t_emb.to(dtype=sample.dtype)
+        emb = self.time_embedding(t_emb)
+
+        if self.addition_embed_type == "text_time":
+            text_embeds = added_cond_kwargs.get("text_embeds")
+            time_ids = added_cond_kwargs.get("time_ids")
+            time_embeds = self.add_time_proj(time_ids.flatten())
+            time_embeds = time_embeds.reshape((text_embeds.shape[0], -1))
+            add_embeds = torch.concat([text_embeds, time_embeds], dim=-1)
+            add_embeds = add_embeds.to(emb.dtype)
+            aug_emb = self.add_embedding(add_embeds)
+            emb = emb + aug_emb
+
+        sample = self.conv_in(sample)
+
+        down_block_res_samples = (sample,)
+        for down_block in self.down_blocks:
+            sample, res_samples = down_block(
+                hidden_states=sample,
+                temb=emb,
+                encoder_hidden_states=encoder_hidden_states,
+            )
+            down_block_res_samples += res_samples
+
+        sample = self.mid_block(sample, emb, encoder_hidden_states=encoder_hidden_states)
+
+        for up_block in self.up_blocks:
+            res_samples = down_block_res_samples[-len(up_block.resnets):]
+            down_block_res_samples = down_block_res_samples[:-len(up_block.resnets)]
+
+            upsample_size = down_block_res_samples[-1].shape[2:] if down_block_res_samples else None
+            sample = up_block(
+                hidden_states=sample,
+                temb=emb,
+                encoder_hidden_states=encoder_hidden_states,
+                res_hidden_states_tuple=res_samples,
+                upsample_size=upsample_size,
+            )
+
+        sample = self.conv_norm_out(sample)
+        sample = self.conv_act(sample)
+        sample = self.conv_out(sample)
+
+        if not return_dict:
+            return (sample,)
+        return sample

diffsynth/models/wan_video_dit.py

@@ -6,6 +6,7 @@ from typing import Tuple, Optional
 from einops import rearrange
 from .wan_video_camera_controller import SimpleAdapter
 from ..core.gradient import gradient_checkpoint_forward
+from .wantodance import WanToDanceRotaryEmbedding, WanToDanceMusicEncoderLayer
 
 try:
     import flash_attn_interface
@@ -283,7 +284,61 @@ class Head(nn.Module):
         return x
 
 
+def wantodance_torch_dfs(model: nn.Module, parent_name='root'):
+    module_names, modules = [], []
+    current_name = parent_name if parent_name else 'root'
+    module_names.append(current_name)
+    modules.append(model)
+    for name, child in model.named_children():
+        if parent_name:
+            child_name = f'{parent_name}.{name}'
+        else:
+            child_name = name
+        child_modules, child_names = wantodance_torch_dfs(child, child_name)
+        module_names += child_names
+        modules += child_modules
+    return modules, module_names
+
+
+class WanToDanceInjector(nn.Module):
+    def __init__(self, all_modules, all_modules_names, dim=2048, num_heads=32, inject_layer=[0, 27]):
+        super().__init__()
+        self.injected_block_id = {}
+        injector_id = 0
+        for mod_name, mod in zip(all_modules_names, all_modules):
+            if isinstance(mod, DiTBlock):
+                for inject_id in inject_layer:
+                    if f'root.transformer_blocks.{inject_id}' == mod_name:
+                        self.injected_block_id[inject_id] = injector_id
+                        injector_id += 1
+
+        self.injector = nn.ModuleList(
+            [
+                CrossAttention(
+                    dim=dim,
+                    num_heads=num_heads,
+                )
+                for _ in range(injector_id)
+            ]
+        )
+        self.injector_pre_norm_feat = nn.ModuleList(
+            [
+                nn.LayerNorm(dim, elementwise_affine=False, eps=1e-6,)
+                for _ in range(injector_id)
+            ]
+        )
+        self.injector_pre_norm_vec = nn.ModuleList(
+            [
+                nn.LayerNorm(dim, elementwise_affine=False, eps=1e-6,)
+                for _ in range(injector_id)
+            ]
+        )
+
+
 class WanModel(torch.nn.Module):
+
+    _repeated_blocks = ["DiTBlock"]
+
     def __init__(
         self,
         dim: int,
@@ -305,6 +360,13 @@ class WanModel(torch.nn.Module):
         require_vae_embedding: bool = True,
         require_clip_embedding: bool = True,
         fuse_vae_embedding_in_latents: bool = False,
+        wantodance_enable_music_inject: bool = False,
+        wantodance_music_inject_layers = [0, 4, 8, 12, 16, 20, 24, 27],
+        wantodance_enable_refimage: bool = False,
+        wantodance_enable_refface: bool = False,
+        wantodance_enable_global: bool = False,
+        wantodance_enable_dynamicfps: bool = False,
+        wantodance_enable_unimodel: bool = False,
     ):
         super().__init__()
         self.dim = dim
@@ -337,7 +399,12 @@ class WanModel(torch.nn.Module):
         ])
         self.head = Head(dim, out_dim, patch_size, eps)
         head_dim = dim // num_heads
-        self.freqs = precompute_freqs_cis_3d(head_dim)
+
+        if wantodance_enable_dynamicfps or wantodance_enable_unimodel:
+            end = int(22350 / 8 + 0.5) # 149f * 30fps * 5s = 22350
+            self.freqs = precompute_freqs_cis_3d(head_dim, end=end)
+        else:
+            self.freqs = precompute_freqs_cis_3d(head_dim)
 
         if has_image_input:
             self.img_emb = MLP(1280, dim, has_pos_emb=has_image_pos_emb)  # clip_feature_dim = 1280
@@ -350,8 +417,83 @@ class WanModel(torch.nn.Module):
         else:
             self.control_adapter = None
 
-    def patchify(self, x: torch.Tensor, control_camera_latents_input: Optional[torch.Tensor] = None):
-        x = self.patch_embedding(x)
+        self.prepare_wantodance(in_dim, dim, num_heads, has_image_pos_emb, out_dim, patch_size, eps,
+                                wantodance_enable_music_inject, wantodance_music_inject_layers, wantodance_enable_refimage, wantodance_enable_refface,
+                                wantodance_enable_global, wantodance_enable_dynamicfps, wantodance_enable_unimodel)
+
+    def prepare_wantodance(
+        self,
+        in_dim, dim, num_heads, has_image_pos_emb, out_dim, patch_size, eps,
+        wantodance_enable_music_inject: bool = False,
+        wantodance_music_inject_layers = [0, 4, 8, 12, 16, 20, 24, 27],
+        wantodance_enable_refimage: bool = False,
+        wantodance_enable_refface: bool = False,
+        wantodance_enable_global: bool = False,
+        wantodance_enable_dynamicfps: bool = False,
+        wantodance_enable_unimodel: bool = False,
+    ):
+        if wantodance_enable_music_inject:
+            all_modules, all_modules_names = wantodance_torch_dfs(self.blocks, parent_name="root.transformer_blocks")
+            self.music_injector = WanToDanceInjector(all_modules, all_modules_names, dim=dim, num_heads=num_heads, inject_layer=wantodance_music_inject_layers)
+        if wantodance_enable_refimage:
+            self.img_emb_refimage = MLP(1280, dim, has_pos_emb=has_image_pos_emb)  # clip_feature_dim = 1280
+        if wantodance_enable_refface:
+            self.img_emb_refface = MLP(1280, dim, has_pos_emb=has_image_pos_emb)  # clip_feature_dim = 1280
+        if wantodance_enable_global or wantodance_enable_dynamicfps or wantodance_enable_unimodel:
+            music_feature_dim = 35
+            ff_size = 1024
+            dropout = 0.1
+            latent_dim = 256
+            nhead = 4
+            activation = F.gelu
+            rotary = WanToDanceRotaryEmbedding(dim=latent_dim)
+            self.music_projection = nn.Linear(music_feature_dim, latent_dim)
+            self.music_encoder = nn.Sequential()
+            for _ in range(2):
+                self.music_encoder.append(
+                    WanToDanceMusicEncoderLayer(
+                        d_model=latent_dim,
+                        nhead=nhead,
+                        dim_feedforward=ff_size,
+                        dropout=dropout,
+                        activation=activation,
+                        batch_first=True,
+                        rotary=rotary,
+                        device='cuda',
+                    )
+                )
+        if wantodance_enable_unimodel:
+            self.patch_embedding_global = nn.Conv3d(in_dim, dim, kernel_size=patch_size, stride=patch_size)
+        if wantodance_enable_unimodel:
+            self.head_global = Head(dim, out_dim, patch_size, eps)
+        self.wantodance_enable_music_inject = wantodance_enable_music_inject
+        self.wantodance_enable_refimage = wantodance_enable_refimage
+        self.wantodance_enable_refface = wantodance_enable_refface
+        self.wantodance_enable_global = wantodance_enable_global
+        self.wantodance_enable_dynamicfps = wantodance_enable_dynamicfps
+        self.wantodance_enable_unimodel = wantodance_enable_unimodel
+
+    def wantodance_after_transformer_block(self, block_idx, hidden_states):
+        if self.wantodance_enable_music_inject:
+            if block_idx in self.music_injector.injected_block_id.keys():
+                audio_attn_id = self.music_injector.injected_block_id[block_idx]
+                audio_emb = self.merged_audio_emb  # b f n c
+                num_frames = audio_emb.shape[1]
+                input_hidden_states = hidden_states.clone()  # b (f h w) c
+                input_hidden_states = rearrange(input_hidden_states, "b (t n) c -> (b t) n c", t=num_frames)
+                attn_hidden_states = self.music_injector.injector_pre_norm_feat[audio_attn_id](input_hidden_states)
+                audio_emb = rearrange(audio_emb, "b t c -> (b t) 1 c", t=num_frames)
+                attn_audio_emb = audio_emb
+                residual_out = self.music_injector.injector[audio_attn_id](attn_hidden_states, attn_audio_emb)
+                residual_out = rearrange(residual_out, "(b t) n c -> b (t n) c", t=num_frames)
+                hidden_states = hidden_states + residual_out
+        return hidden_states
+
+    def patchify(self, x: torch.Tensor, control_camera_latents_input: Optional[torch.Tensor] = None, enable_wantodance_global=False):
+        if enable_wantodance_global:
+            x = self.patch_embedding_global(x)
+        else:
+            x = self.patch_embedding(x)
         if self.control_adapter is not None and control_camera_latents_input is not None:
             y_camera = self.control_adapter(control_camera_latents_input)
             x = [u + v for u, v in zip(x, y_camera)]

diffsynth/models/wan_video_vae.py

@@ -1247,6 +1247,22 @@ class WanVideoVAE(nn.Module):
         return videos
 
 
+    def encode_framewise(self, videos, device):
+        hidden_states = []
+        for i in range(videos.shape[2]):
+            hidden_states.append(self.single_encode(videos[:, :, i:i+1], device))
+        hidden_states = torch.concat(hidden_states, dim=2)
+        return hidden_states
+    
+
+    def decode_framewise(self, hidden_states, device):
+        video = []
+        for i in range(hidden_states.shape[2]):
+            video.append(self.single_decode(hidden_states[:, :, i:i+1], device))
+        video = torch.concat(video, dim=2)
+        return video
+
+
     @staticmethod
     def state_dict_converter():
         return WanVideoVAEStateDictConverter()

diffsynth/models/wantodance.py

@@ -0,0 +1,209 @@
+from inspect import isfunction
+from math import log, pi
+
+import torch
+from einops import rearrange, repeat
+from torch import einsum, nn
+
+from typing import Any, Callable, List, Optional, Union
+from torch import Tensor
+import torch.nn.functional as F
+
+# helper functions
+
+
+def exists(val):
+    return val is not None
+
+
+def broadcat(tensors, dim=-1):
+    num_tensors = len(tensors)
+    shape_lens = set(list(map(lambda t: len(t.shape), tensors)))
+    assert len(shape_lens) == 1, "tensors must all have the same number of dimensions"
+    shape_len = list(shape_lens)[0]
+
+    dim = (dim + shape_len) if dim < 0 else dim
+    dims = list(zip(*map(lambda t: list(t.shape), tensors)))
+
+    expandable_dims = [(i, val) for i, val in enumerate(dims) if i != dim]
+    assert all(
+        [*map(lambda t: len(set(t[1])) <= 2, expandable_dims)]
+    ), "invalid dimensions for broadcastable concatentation"
+    max_dims = list(map(lambda t: (t[0], max(t[1])), expandable_dims))
+    expanded_dims = list(map(lambda t: (t[0], (t[1],) * num_tensors), max_dims))
+    expanded_dims.insert(dim, (dim, dims[dim]))
+    expandable_shapes = list(zip(*map(lambda t: t[1], expanded_dims)))
+    tensors = list(map(lambda t: t[0].expand(*t[1]), zip(tensors, expandable_shapes)))
+    return torch.cat(tensors, dim=dim)
+
+
+# rotary embedding helper functions
+
+
+def rotate_half(x):
+    x = rearrange(x, "... (d r) -> ... d r", r=2)
+    x1, x2 = x.unbind(dim=-1)
+    x = torch.stack((-x2, x1), dim=-1)
+    return rearrange(x, "... d r -> ... (d r)")
+
+
+def apply_rotary_emb(freqs, t, start_index=0):
+    freqs = freqs.to(t)
+    rot_dim = freqs.shape[-1]
+    end_index = start_index + rot_dim
+    assert (
+        rot_dim <= t.shape[-1]
+    ), f"feature dimension {t.shape[-1]} is not of sufficient size to rotate in all the positions {rot_dim}"
+    t_left, t, t_right = (
+        t[..., :start_index],
+        t[..., start_index:end_index],
+        t[..., end_index:],
+    )
+    t = (t * freqs.cos()) + (rotate_half(t) * freqs.sin())
+    return torch.cat((t_left, t, t_right), dim=-1)
+
+
+# learned rotation helpers
+
+
+def apply_learned_rotations(rotations, t, start_index=0, freq_ranges=None):
+    if exists(freq_ranges):
+        rotations = einsum("..., f -> ... f", rotations, freq_ranges)
+        rotations = rearrange(rotations, "... r f -> ... (r f)")
+
+    rotations = repeat(rotations, "... n -> ... (n r)", r=2)
+    return apply_rotary_emb(rotations, t, start_index=start_index)
+
+
+# classes
+
+
+class WanToDanceRotaryEmbedding(nn.Module):
+    def __init__(
+        self,
+        dim,
+        custom_freqs=None,
+        freqs_for="lang",
+        theta=10000,
+        max_freq=10,
+        num_freqs=1,
+        learned_freq=False,
+    ):
+        super().__init__()
+        if exists(custom_freqs):
+            freqs = custom_freqs
+        elif freqs_for == "lang":
+            freqs = 1.0 / (
+                theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim)
+            )
+        elif freqs_for == "pixel":
+            freqs = torch.linspace(1.0, max_freq / 2, dim // 2) * pi
+        elif freqs_for == "constant":
+            freqs = torch.ones(num_freqs).float()
+        else:
+            raise ValueError(f"unknown modality {freqs_for}")
+
+        self.cache = dict()
+
+        if learned_freq:
+            self.freqs = nn.Parameter(freqs)
+        else:
+            self.register_buffer("freqs", freqs, persistent=False)
+
+    def rotate_queries_or_keys(self, t, seq_dim=-2):
+        device = t.device
+        seq_len = t.shape[seq_dim]
+        freqs = self.forward(
+            lambda: torch.arange(seq_len, device=device), cache_key=seq_len
+        )
+        return apply_rotary_emb(freqs, t)
+
+    def forward(self, t, cache_key=None):
+        if exists(cache_key) and cache_key in self.cache:
+            return self.cache[cache_key]
+
+        if isfunction(t):
+            t = t()
+
+        # freqs = self.freqs
+        freqs = self.freqs.to(t.device)
+
+        freqs = torch.einsum("..., f -> ... f", t.type(freqs.dtype), freqs)
+        freqs = repeat(freqs, "... n -> ... (n r)", r=2)
+
+        if exists(cache_key):
+            self.cache[cache_key] = freqs
+
+        return freqs
+
+
+class WanToDanceMusicEncoderLayer(nn.Module):
+    def __init__(
+        self,
+        d_model: int,
+        nhead: int,
+        dim_feedforward: int = 2048,
+        dropout: float = 0.1,
+        activation: Union[str, Callable[[Tensor], Tensor]] = F.relu,
+        layer_norm_eps: float = 1e-5,
+        batch_first: bool = False,
+        norm_first: bool = True,
+        device=None,
+        dtype=None,
+        rotary=None,
+    ) -> None:
+        super().__init__()
+        self.self_attn = nn.MultiheadAttention(
+            d_model, nhead, dropout=dropout, batch_first=batch_first, device=device, dtype=dtype
+        )
+        # Implementation of Feedforward model
+        self.linear1 = nn.Linear(d_model, dim_feedforward)
+        self.dropout = nn.Dropout(dropout)
+        self.linear2 = nn.Linear(dim_feedforward, d_model)
+
+        self.norm_first = norm_first
+        self.norm1 = nn.LayerNorm(d_model, eps=layer_norm_eps)
+        self.norm2 = nn.LayerNorm(d_model, eps=layer_norm_eps)
+        self.dropout1 = nn.Dropout(dropout)
+        self.dropout2 = nn.Dropout(dropout)
+        self.activation = activation
+
+        self.rotary = rotary
+        self.use_rotary = rotary is not None
+
+    # self-attention block
+    def _sa_block(
+        self, x: Tensor, attn_mask: Optional[Tensor], key_padding_mask: Optional[Tensor]
+    ) -> Tensor:
+        qk = self.rotary.rotate_queries_or_keys(x) if self.use_rotary else x
+        x = self.self_attn(
+            qk,
+            qk,
+            x,
+            attn_mask=attn_mask,
+            key_padding_mask=key_padding_mask,
+            need_weights=False,
+        )[0]
+        return self.dropout1(x)
+
+    # feed forward block
+    def _ff_block(self, x: Tensor) -> Tensor:
+        x = self.linear2(self.dropout(self.activation(self.linear1(x))))
+        return self.dropout2(x)
+
+    def forward(
+        self,
+        src: Tensor,
+        src_mask: Optional[Tensor] = None,
+        src_key_padding_mask: Optional[Tensor] = None,
+    ) -> Tensor:
+        x = src
+        if self.norm_first:
+            self.norm1.to(device=x.device)
+            self.norm2.to(device=x.device)
+            x = x + self._sa_block(self.norm1(x), src_mask, src_key_padding_mask)
+            x = x + self._ff_block(self.norm2(x))
+        else:
+            x = self.norm1(x + self._sa_block(x, src_mask, src_key_padding_mask))
+            x = self.norm2(x + self._ff_block(x))
+        return x

diffsynth/models/z_image_dit.py

@@ -326,6 +326,7 @@ class RopeEmbedder:
 class ZImageDiT(nn.Module):
     _supports_gradient_checkpointing = True
     _no_split_modules = ["ZImageTransformerBlock"]
+    _repeated_blocks = ["ZImageTransformerBlock"]
 
     def __init__(
         self,

diffsynth/pipelines/anima_image.py

@@ -39,6 +39,7 @@ class AnimaImagePipeline(BasePipeline):
             AnimaUnit_PromptEmbedder(),
         ]
         self.model_fn = model_fn_anima
+        self.compilable_models = ["dit"]
     
     
     @staticmethod

diffsynth/pipelines/ernie_image.py

@@ -0,0 +1,266 @@
+"""
+ERNIE-Image Text-to-Image Pipeline for DiffSynth-Studio.
+
+Architecture: SharedAdaLN DiT + RoPE 3D + Joint Image-Text Attention.
+"""
+
+import torch
+from typing import Union, Optional
+from tqdm import tqdm
+from transformers import AutoTokenizer
+
+from ..core.device.npu_compatible_device import get_device_type
+from ..diffusion import FlowMatchScheduler
+from ..core import ModelConfig
+from ..diffusion.base_pipeline import BasePipeline, PipelineUnit
+from ..models.ernie_image_text_encoder import ErnieImageTextEncoder
+from ..models.ernie_image_dit import ErnieImageDiT
+from ..models.flux2_vae import Flux2VAE
+
+
+class ErnieImagePipeline(BasePipeline):
+
+    def __init__(self, device=get_device_type(), torch_dtype=torch.bfloat16):
+        super().__init__(
+            device=device, torch_dtype=torch_dtype,
+            height_division_factor=16, width_division_factor=16,
+        )
+        self.scheduler = FlowMatchScheduler("ERNIE-Image")
+        self.text_encoder: ErnieImageTextEncoder = None
+        self.dit: ErnieImageDiT = None
+        self.vae: Flux2VAE = None
+        self.tokenizer: AutoTokenizer = None
+
+        self.in_iteration_models = ("dit",)
+        self.units = [
+            ErnieImageUnit_ShapeChecker(),
+            ErnieImageUnit_PromptEmbedder(),
+            ErnieImageUnit_NoiseInitializer(),
+            ErnieImageUnit_InputImageEmbedder(),
+        ]
+        self.model_fn = model_fn_ernie_image
+        self.compilable_models = ["dit"]
+
+    @staticmethod
+    def from_pretrained(
+        torch_dtype: torch.dtype = torch.bfloat16,
+        device: Union[str, torch.device] = get_device_type(),
+        model_configs: list[ModelConfig] = [],
+        tokenizer_config: ModelConfig = ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+        vram_limit: float = None,
+    ):
+        pipe = ErnieImagePipeline(device=device, torch_dtype=torch_dtype)
+        model_pool = pipe.download_and_load_models(model_configs, vram_limit)
+
+        pipe.text_encoder = model_pool.fetch_model("ernie_image_text_encoder")
+        pipe.dit = model_pool.fetch_model("ernie_image_dit")
+        pipe.vae = model_pool.fetch_model("flux2_vae")
+
+        if tokenizer_config is not None:
+            tokenizer_config.download_if_necessary()
+            pipe.tokenizer = AutoTokenizer.from_pretrained(tokenizer_config.path)
+
+        pipe.vram_management_enabled = pipe.check_vram_management_state()
+        return pipe
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        # Prompt
+        prompt: str,
+        negative_prompt: str = "",
+        cfg_scale: float = 4.0,
+        # Shape
+        height: int = 1024,
+        width: int = 1024,
+        # Randomness
+        seed: int = None,
+        rand_device: str = "cuda",
+        # Steps
+        num_inference_steps: int = 50,
+        sigma_shift: float = 3.0,
+        # Progress bar
+        progress_bar_cmd=tqdm,
+    ):
+        # Scheduler
+        self.scheduler.set_timesteps(num_inference_steps=num_inference_steps, shift=sigma_shift)
+
+        # Parameters
+        inputs_posi = {"prompt": prompt}
+        inputs_nega = {"negative_prompt": negative_prompt}
+        inputs_shared = {
+            "height": height, "width": width, "seed": seed,
+            "cfg_scale": cfg_scale, "num_inference_steps": num_inference_steps,
+            "rand_device": rand_device,
+        }
+        for unit in self.units:
+            inputs_shared, inputs_posi, inputs_nega = self.unit_runner(unit, self, inputs_shared, inputs_posi, inputs_nega)
+
+        # Denoise
+        self.load_models_to_device(self.in_iteration_models)
+        models = {name: getattr(self, name) for name in self.in_iteration_models}
+        for progress_id, timestep in enumerate(progress_bar_cmd(self.scheduler.timesteps)):
+            timestep = timestep.unsqueeze(0).to(dtype=self.torch_dtype, device=self.device)
+            noise_pred = self.cfg_guided_model_fn(
+                self.model_fn, cfg_scale,
+                inputs_shared, inputs_posi, inputs_nega,
+                **models, timestep=timestep, progress_id=progress_id
+            )
+            inputs_shared["latents"] = self.step(self.scheduler, progress_id=progress_id, noise_pred=noise_pred, **inputs_shared)
+
+        # Decode
+        self.load_models_to_device(['vae'])
+        latents = inputs_shared["latents"]
+        image = self.vae.decode(latents)
+        image = self.vae_output_to_image(image)
+        self.load_models_to_device([])
+        return image
+
+
+class ErnieImageUnit_ShapeChecker(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width"),
+            output_params=("height", "width"),
+        )
+
+    def process(self, pipe: ErnieImagePipeline, height, width):
+        height, width = pipe.check_resize_height_width(height, width)
+        return {"height": height, "width": width}
+
+
+class ErnieImageUnit_PromptEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            seperate_cfg=True,
+            input_params_posi={"prompt": "prompt"},
+            input_params_nega={"prompt": "negative_prompt"},
+            output_params=("prompt_embeds", "prompt_embeds_mask"),
+            onload_model_names=("text_encoder",)
+        )
+
+    def encode_prompt(self, pipe: ErnieImagePipeline, prompt):
+        if isinstance(prompt, str):
+            prompt = [prompt]
+
+        text_hiddens = []
+        text_lens_list = []
+        for p in prompt:
+            ids = pipe.tokenizer(
+                p,
+                add_special_tokens=True,
+                truncation=True,
+                padding=False,
+            )["input_ids"]
+
+            if len(ids) == 0:
+                if pipe.tokenizer.bos_token_id is not None:
+                    ids = [pipe.tokenizer.bos_token_id]
+                else:
+                    ids = [0]
+
+            input_ids = torch.tensor([ids], device=pipe.device)
+            outputs = pipe.text_encoder(
+                input_ids=input_ids,
+            )
+            # Text encoder returns tuple of (hidden_states_tuple,) where each layer's hidden state is included
+            all_hidden_states = outputs[0]
+            hidden = all_hidden_states[-2][0]  # [T, H] - second to last layer
+            text_hiddens.append(hidden)
+            text_lens_list.append(hidden.shape[0])
+
+        # Pad to uniform length
+        if len(text_hiddens) == 0:
+            text_in_dim = pipe.text_encoder.config.hidden_size if hasattr(pipe.text_encoder, 'config') else 3072
+            return {
+                "prompt_embeds": torch.zeros((0, 0, text_in_dim), device=pipe.device, dtype=pipe.torch_dtype),
+                "prompt_embeds_mask": torch.zeros((0,), device=pipe.device, dtype=torch.long),
+            }
+
+        normalized = [th.to(pipe.device).to(pipe.torch_dtype) for th in text_hiddens]
+        text_lens = torch.tensor([t.shape[0] for t in normalized], device=pipe.device, dtype=torch.long)
+        Tmax = int(text_lens.max().item())
+        text_in_dim = normalized[0].shape[1]
+        text_bth = torch.zeros((len(normalized), Tmax, text_in_dim), device=pipe.device, dtype=pipe.torch_dtype)
+        for i, t in enumerate(normalized):
+            text_bth[i, :t.shape[0], :] = t
+
+        return {"prompt_embeds": text_bth, "prompt_embeds_mask": text_lens}
+
+    def process(self, pipe: ErnieImagePipeline, prompt):
+        pipe.load_models_to_device(self.onload_model_names)
+        if pipe.text_encoder is not None:
+            return self.encode_prompt(pipe, prompt)
+        return {}
+
+
+class ErnieImageUnit_NoiseInitializer(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width", "seed", "rand_device"),
+            output_params=("noise",),
+        )
+
+    def process(self, pipe: ErnieImagePipeline, height, width, seed, rand_device):
+        latent_h = height // pipe.height_division_factor
+        latent_w = width // pipe.width_division_factor
+        latent_channels = pipe.dit.in_channels
+
+        # Use pipeline device if rand_device is not specified
+        if rand_device is None:
+            rand_device = str(pipe.device)
+
+        noise = pipe.generate_noise(
+            (1, latent_channels, latent_h, latent_w),
+            seed=seed,
+            rand_device=rand_device,
+            rand_torch_dtype=pipe.torch_dtype,
+        )
+        return {"noise": noise}
+
+
+class ErnieImageUnit_InputImageEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("input_image", "noise"),
+            output_params=("latents", "input_latents"),
+            onload_model_names=("vae",)
+        )
+
+    def process(self, pipe: ErnieImagePipeline, input_image, noise):
+        if input_image is None:
+            # T2I path: use noise directly as initial latents
+            return {"latents": noise, "input_latents": None}
+
+        # I2I path: VAE encode input image
+        pipe.load_models_to_device(['vae'])
+        image = pipe.preprocess_image(input_image).to(device=pipe.device, dtype=pipe.torch_dtype)
+        input_latents = pipe.vae.encode(image)
+
+        if pipe.scheduler.training:
+            return {"latents": noise, "input_latents": input_latents}
+        else:
+            # In inference mode, add noise to encoded latents
+            latents = pipe.scheduler.add_noise(input_latents, noise, timestep=pipe.scheduler.timesteps[0])
+            return {"latents": latents}
+
+
+def model_fn_ernie_image(
+    dit: ErnieImageDiT,
+    latents=None,
+    timestep=None,
+    prompt_embeds=None,
+    prompt_embeds_mask=None,
+    use_gradient_checkpointing=False,
+    use_gradient_checkpointing_offload=False,
+    **kwargs,
+):
+    output = dit(
+        hidden_states=latents,
+        timestep=timestep,
+        text_bth=prompt_embeds,
+        text_lens=prompt_embeds_mask,
+        use_gradient_checkpointing=use_gradient_checkpointing,
+        use_gradient_checkpointing_offload=use_gradient_checkpointing_offload,
+    )
+    return output

diffsynth/pipelines/flux2_image.py

@@ -42,6 +42,7 @@ class Flux2ImagePipeline(BasePipeline):
             Flux2Unit_ImageIDs(),
         ]
         self.model_fn = model_fn_flux2
+        self.compilable_models = ["dit"]
     
     
     @staticmethod

diffsynth/pipelines/flux_image.py

@@ -103,6 +103,7 @@ class FluxImagePipeline(BasePipeline):
             FluxImageUnit_LoRAEncode(),
         ]
         self.model_fn = model_fn_flux_image
+        self.compilable_models = ["dit"]
         self.lora_loader = FluxLoRALoader
 
     def enable_lora_merger(self):

diffsynth/pipelines/joyai_image.py

@@ -0,0 +1,282 @@
+import torch
+from PIL import Image
+from typing import Union, Optional
+from tqdm import tqdm
+from einops import rearrange
+
+from ..core.device.npu_compatible_device import get_device_type
+from ..diffusion import FlowMatchScheduler
+from ..core import ModelConfig
+from ..diffusion.base_pipeline import BasePipeline, PipelineUnit
+from ..models.joyai_image_dit import JoyAIImageDiT
+from ..models.joyai_image_text_encoder import JoyAIImageTextEncoder
+from ..models.wan_video_vae import WanVideoVAE
+
+class JoyAIImagePipeline(BasePipeline):
+
+    def __init__(self, device=get_device_type(), torch_dtype=torch.bfloat16):
+        super().__init__(
+            device=device, torch_dtype=torch_dtype,
+            height_division_factor=16, width_division_factor=16,
+        )
+        self.scheduler = FlowMatchScheduler("Wan")
+        self.text_encoder: JoyAIImageTextEncoder = None
+        self.dit: JoyAIImageDiT = None
+        self.vae: WanVideoVAE = None
+        self.processor = None
+        self.in_iteration_models = ("dit",)
+
+        self.units = [
+            JoyAIImageUnit_ShapeChecker(),
+            JoyAIImageUnit_EditImageEmbedder(),
+            JoyAIImageUnit_PromptEmbedder(),
+            JoyAIImageUnit_NoiseInitializer(),
+            JoyAIImageUnit_InputImageEmbedder(),
+        ]
+        self.model_fn = model_fn_joyai_image
+        self.compilable_models = ["dit"]
+
+    @staticmethod
+    def from_pretrained(
+        torch_dtype: torch.dtype = torch.bfloat16,
+        device: Union[str, torch.device] = get_device_type(),
+        model_configs: list[ModelConfig] = [],
+        # Processor
+        processor_config: ModelConfig = None,
+        # Optional
+        vram_limit: float = None,
+    ):
+        pipe = JoyAIImagePipeline(device=device, torch_dtype=torch_dtype)
+        model_pool = pipe.download_and_load_models(model_configs, vram_limit)
+
+        pipe.text_encoder = model_pool.fetch_model("joyai_image_text_encoder")
+        pipe.dit = model_pool.fetch_model("joyai_image_dit")
+        pipe.vae = model_pool.fetch_model("wan_video_vae")
+
+        if processor_config is not None:
+            processor_config.download_if_necessary()
+            from transformers import AutoProcessor
+            pipe.processor = AutoProcessor.from_pretrained(processor_config.path)
+
+        pipe.vram_management_enabled = pipe.check_vram_management_state()
+        return pipe
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        # Prompt
+        prompt: str,
+        negative_prompt: str = "",
+        cfg_scale: float = 5.0,
+        # Image
+        edit_image: Image.Image = None,
+        denoising_strength: float = 1.0,
+        # Shape
+        height: int = 1024,
+        width: int = 1024,
+        # Randomness
+        seed: int = None,
+        # Steps
+        max_sequence_length: int = 4096,
+        num_inference_steps: int = 30,
+        # Tiling
+        tiled: Optional[bool] = False,
+        tile_size: Optional[tuple[int, int]] = (30, 52),
+        tile_stride: Optional[tuple[int, int]] = (15, 26),
+        # Scheduler
+        shift: Optional[float] = 4.0,
+        # Progress bar
+        progress_bar_cmd=tqdm,
+    ):
+        # Scheduler
+        self.scheduler.set_timesteps(num_inference_steps, denoising_strength=denoising_strength, shift=shift)
+
+        # Parameters
+        inputs_posi = {"prompt": prompt}
+        inputs_nega = {"negative_prompt": negative_prompt}
+        inputs_shared = {
+            "cfg_scale": cfg_scale,
+            "edit_image": edit_image,
+            "denoising_strength": denoising_strength,
+            "height": height, "width": width,
+            "seed": seed, "max_sequence_length": max_sequence_length,
+            "tiled": tiled, "tile_size": tile_size, "tile_stride": tile_stride,
+        }
+
+        # Unit chain
+        for unit in self.units:
+            inputs_shared, inputs_posi, inputs_nega = self.unit_runner(
+                unit, self, inputs_shared, inputs_posi, inputs_nega
+            )
+
+        # Denoise
+        self.load_models_to_device(self.in_iteration_models)
+        models = {name: getattr(self, name) for name in self.in_iteration_models}
+        for progress_id, timestep in enumerate(progress_bar_cmd(self.scheduler.timesteps)):
+            timestep = timestep.unsqueeze(0).to(dtype=self.torch_dtype, device=self.device)
+            noise_pred = self.cfg_guided_model_fn(
+                self.model_fn, cfg_scale,
+                inputs_shared, inputs_posi, inputs_nega,
+                **models, timestep=timestep, progress_id=progress_id
+            )
+            inputs_shared["latents"] = self.step(self.scheduler, progress_id=progress_id, noise_pred=noise_pred, **inputs_shared)
+
+        # Decode
+        self.load_models_to_device(['vae'])
+        latents = rearrange(inputs_shared["latents"], "b n c f h w -> (b n) c f h w")
+        image = self.vae.decode(latents, device=self.device)[0]
+        image = self.vae_output_to_image(image, pattern="C 1 H W")
+        self.load_models_to_device([])
+        return image
+
+
+class JoyAIImageUnit_ShapeChecker(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width"),
+            output_params=("height", "width"),
+        )
+
+    def process(self, pipe: "JoyAIImagePipeline", height, width):
+        height, width = pipe.check_resize_height_width(height, width)
+        return {"height": height, "width": width}
+
+
+class JoyAIImageUnit_PromptEmbedder(PipelineUnit):
+    prompt_template_encode = {
+        'image':
+            "<|im_start|>system\n \\nDescribe the image by detailing the color, shape, size, texture, quantity, text, spatial relationships of the objects and background:<|im_end|>\n<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n",
+        'multiple_images':
+            "<|im_start|>system\n \\nDescribe the image by detailing the color, shape, size, texture, quantity, text, spatial relationships of the objects and background:<|im_end|>\n{}<|im_start|>assistant\n",
+        'video':
+            "<|im_start|>system\n \\nDescribe the video by detailing the following aspects:\n1. The main content and theme of the video.\n2. The color, shape, size, texture, quantity, text, and spatial relationships of the objects.\n3. Actions, events, behaviors temporal relationships, physical movement changes of the objects.\n4. background environment, light, style and atmosphere.\n5. camera angles, movements, and transitions used in the video:<|im_end|>\n<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n"
+    }
+    prompt_template_encode_start_idx = {'image': 34, 'multiple_images': 34, 'video': 91}
+    def __init__(self):
+        super().__init__(
+            seperate_cfg=True,
+            input_params_posi={"prompt": "prompt", "positive": "positive"},
+            input_params_nega={"prompt": "negative_prompt", "positive": "positive"},
+            input_params=("edit_image", "max_sequence_length"),
+            output_params=("prompt_embeds", "prompt_embeds_mask"),
+            onload_model_names=("joyai_image_text_encoder",),
+        )
+
+    def process(self, pipe: "JoyAIImagePipeline", prompt, positive, edit_image, max_sequence_length):
+        pipe.load_models_to_device(self.onload_model_names)
+
+        has_image = edit_image is not None
+
+        if has_image:
+            prompt_embeds, prompt_embeds_mask = self._encode_with_image(pipe, prompt, edit_image, max_sequence_length)
+        else:
+            prompt_embeds, prompt_embeds_mask = self._encode_text_only(pipe, prompt, max_sequence_length)
+
+        return {"prompt_embeds": prompt_embeds, "prompt_embeds_mask": prompt_embeds_mask}
+
+    def _encode_with_image(self, pipe, prompt, edit_image, max_sequence_length):
+        template = self.prompt_template_encode['multiple_images']
+        drop_idx = self.prompt_template_encode_start_idx['multiple_images']
+
+        image_tokens = '<image>\n'
+        prompt = f"<|im_start|>user\n{image_tokens}{prompt}<|im_end|>\n"
+        prompt = prompt.replace('<image>\n', '<|vision_start|><|image_pad|><|vision_end|>')
+        prompt = template.format(prompt)
+        inputs = pipe.processor(text=[prompt], images=[edit_image], padding=True, return_tensors="pt").to(pipe.device)
+        last_hidden_states = pipe.text_encoder(**inputs)
+
+        prompt_embeds = last_hidden_states[:, drop_idx:]
+        prompt_embeds_mask = inputs['attention_mask'][:, drop_idx:]
+
+        if max_sequence_length is not None and prompt_embeds.shape[1] > max_sequence_length:
+            prompt_embeds = prompt_embeds[:, -max_sequence_length:, :]
+            prompt_embeds_mask = prompt_embeds_mask[:, -max_sequence_length:]
+
+        return prompt_embeds, prompt_embeds_mask
+
+    def _encode_text_only(self, pipe, prompt, max_sequence_length):
+        # TODO: may support for text-only encoding in the future.
+        raise NotImplementedError("Text-only encoding is not implemented yet. Please provide edit_image for now.")
+        return prompt_embeds, encoder_attention_mask
+
+
+class JoyAIImageUnit_EditImageEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("edit_image", "tiled", "tile_size", "tile_stride", "height", "width"),
+            output_params=("ref_latents", "num_items", "is_multi_item"),
+            onload_model_names=("wan_video_vae",),
+        )
+
+    def process(self, pipe: "JoyAIImagePipeline", edit_image, tiled, tile_size, tile_stride, height, width):
+        if edit_image is None:
+            return {}
+        pipe.load_models_to_device(self.onload_model_names)
+        # Resize edit image to match target dimensions (from ShapeChecker) to ensure ref_latents matches latents
+        edit_image = edit_image.resize((width, height), Image.LANCZOS)
+        images = [pipe.preprocess_image(edit_image).transpose(0, 1)]
+        latents = pipe.vae.encode(images, device=pipe.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride)
+        ref_vae = rearrange(latents, "(b n) c 1 h w -> b n c 1 h w", n=1).to(device=pipe.device, dtype=pipe.torch_dtype)
+
+        return {"ref_latents": ref_vae, "edit_image": edit_image}
+
+
+class JoyAIImageUnit_NoiseInitializer(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("seed", "height", "width", "rand_device"),
+            output_params=("noise"),
+        )
+
+    def process(self, pipe: "JoyAIImagePipeline", seed, height, width, rand_device):
+        latent_h = height // pipe.vae.upsampling_factor
+        latent_w = width // pipe.vae.upsampling_factor
+        shape = (1, 1, pipe.vae.z_dim, 1, latent_h, latent_w)
+        noise = pipe.generate_noise(shape, seed=seed, rand_device=rand_device, rand_torch_dtype=pipe.torch_dtype)
+        return {"noise": noise}
+
+
+class JoyAIImageUnit_InputImageEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("input_image", "noise", "tiled", "tile_size", "tile_stride"),
+            output_params=("latents", "input_latents"),
+            onload_model_names=("vae",),
+        )
+
+    def process(self, pipe: JoyAIImagePipeline, input_image, noise, tiled, tile_size, tile_stride):
+        if input_image is None:
+            return {"latents": noise}
+        pipe.load_models_to_device(self.onload_model_names)
+        if isinstance(input_image, Image.Image):
+            input_image = [input_image]
+        input_image = [pipe.preprocess_image(img).transpose(0, 1) for img in input_image]
+        latents = pipe.vae.encode(input_image, device=pipe.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride)
+        input_latents = rearrange(latents, "(b n) c 1 h w -> b n c 1 h w", n=(len(input_image)))
+        return {"latents": noise, "input_latents": input_latents}
+
+def model_fn_joyai_image(
+    dit,
+    latents,
+    timestep,
+    prompt_embeds,
+    prompt_embeds_mask,
+    ref_latents=None,
+    use_gradient_checkpointing=False,
+    use_gradient_checkpointing_offload=False,
+    **kwargs,
+):
+
+    img = torch.cat([ref_latents, latents], dim=1) if ref_latents is not None else latents
+
+    img = dit(
+        hidden_states=img,
+        timestep=timestep,
+        encoder_hidden_states=prompt_embeds,
+        encoder_hidden_states_mask=prompt_embeds_mask,
+        use_gradient_checkpointing=use_gradient_checkpointing,
+        use_gradient_checkpointing_offload=use_gradient_checkpointing_offload,
+    )
+
+    img = img[:, -latents.size(1):]
+    return img

diffsynth/pipelines/ltx2_audio_video.py

@@ -76,6 +76,7 @@ class LTX2AudioVideoPipeline(BasePipeline):
             LTX2AudioVideoUnit_SetScheduleStage2(),
         ]
         self.model_fn = model_fn_ltx2
+        self.compilable_models = ["dit"]
 
         self.default_negative_prompt = {
             "LTX-2": (

diffsynth/pipelines/mova_audio_video.py

@@ -52,6 +52,7 @@ class MovaAudioVideoPipeline(BasePipeline):
             MovaAudioVideoUnit_UnifiedSequenceParallel(),
         ]
         self.model_fn = model_fn_mova_audio_video
+        self.compilable_models = ["video_dit", "video_dit2", "audio_dit"]
 
     def enable_usp(self):
         from ..utils.xfuser import get_sequence_parallel_world_size, usp_attn_forward

diffsynth/pipelines/qwen_image.py

@@ -56,6 +56,7 @@ class QwenImagePipeline(BasePipeline):
             QwenImageUnit_BlockwiseControlNet(),
         ]
         self.model_fn = model_fn_qwen_image
+        self.compilable_models = ["dit"]
     
     
     @staticmethod

diffsynth/pipelines/stable_diffusion.py

@@ -0,0 +1,230 @@
+import torch
+from PIL import Image
+from tqdm import tqdm
+from typing import Union
+
+from ..core.device.npu_compatible_device import get_device_type
+from ..diffusion.ddim_scheduler import DDIMScheduler
+from ..core import ModelConfig
+from ..diffusion.base_pipeline import BasePipeline, PipelineUnit
+
+from transformers import AutoTokenizer, CLIPTextModel
+from ..models.stable_diffusion_text_encoder import SDTextEncoder
+from ..models.stable_diffusion_unet import UNet2DConditionModel
+from ..models.stable_diffusion_vae import StableDiffusionVAE
+
+
+class StableDiffusionPipeline(BasePipeline):
+
+    def __init__(self, device=get_device_type(), torch_dtype=torch.float16):
+        super().__init__(
+            device=device, torch_dtype=torch_dtype,
+            height_division_factor=8, width_division_factor=8,
+        )
+        self.scheduler = DDIMScheduler()
+        self.text_encoder: SDTextEncoder = None
+        self.unet: UNet2DConditionModel = None
+        self.vae: StableDiffusionVAE = None
+        self.tokenizer: AutoTokenizer = None
+
+        self.in_iteration_models = ("unet",)
+        self.units = [
+            SDUnit_ShapeChecker(),
+            SDUnit_PromptEmbedder(),
+            SDUnit_NoiseInitializer(),
+            SDUnit_InputImageEmbedder(),
+        ]
+        self.model_fn = model_fn_stable_diffusion
+        self.compilable_models = ["unet"]
+
+    @staticmethod
+    def from_pretrained(
+        torch_dtype: torch.dtype = torch.float16,
+        device: Union[str, torch.device] = get_device_type(),
+        model_configs: list[ModelConfig] = [],
+        tokenizer_config: ModelConfig = None,
+        vram_limit: float = None,
+    ):
+        pipe = StableDiffusionPipeline(device=device, torch_dtype=torch_dtype)
+        # Override vram_config to use the specified torch_dtype for all models
+        for mc in model_configs:
+            mc._vram_config_override = {
+                'onload_dtype': torch_dtype,
+                'computation_dtype': torch_dtype,
+            }
+        model_pool = pipe.download_and_load_models(model_configs, vram_limit)
+        pipe.text_encoder = model_pool.fetch_model("stable_diffusion_text_encoder")
+        pipe.unet = model_pool.fetch_model("stable_diffusion_unet")
+        pipe.vae = model_pool.fetch_model("stable_diffusion_vae")
+        if tokenizer_config is not None:
+            tokenizer_config.download_if_necessary()
+            pipe.tokenizer = AutoTokenizer.from_pretrained(tokenizer_config.path)
+        pipe.vram_management_enabled = pipe.check_vram_management_state()
+        return pipe
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        prompt: str,
+        negative_prompt: str = "",
+        cfg_scale: float = 7.5,
+        height: int = 512,
+        width: int = 512,
+        seed: int = None,
+        rand_device: str = "cpu",
+        num_inference_steps: int = 50,
+        eta: float = 0.0,
+        guidance_rescale: float = 0.0,
+        progress_bar_cmd=tqdm,
+    ):
+        # 1. Scheduler
+        self.scheduler.set_timesteps(
+            num_inference_steps, eta=eta,
+        )
+
+        # 2. Three-dict input preparation
+        inputs_posi = {"prompt": prompt}
+        inputs_nega = {"negative_prompt": negative_prompt}
+        inputs_shared = {
+            "cfg_scale": cfg_scale,
+            "height": height, "width": width,
+            "seed": seed, "rand_device": rand_device,
+            "guidance_rescale": guidance_rescale,
+        }
+
+        # 3. Unit chain execution
+        for unit in self.units:
+            inputs_shared, inputs_posi, inputs_nega = self.unit_runner(
+                unit, self, inputs_shared, inputs_posi, inputs_nega
+            )
+
+        # 4. Denoise loop
+        self.load_models_to_device(self.in_iteration_models)
+        models = {name: getattr(self, name) for name in self.in_iteration_models}
+        for progress_id, timestep in enumerate(progress_bar_cmd(self.scheduler.timesteps)):
+            timestep = timestep.unsqueeze(0).to(dtype=self.torch_dtype, device=self.device)
+            noise_pred = self.cfg_guided_model_fn(
+                self.model_fn, cfg_scale,
+                inputs_shared, inputs_posi, inputs_nega,
+                **models, timestep=timestep, progress_id=progress_id
+            )
+            inputs_shared["latents"] = self.step(
+                self.scheduler, progress_id=progress_id, noise_pred=noise_pred, **inputs_shared
+            )
+
+        # 5. VAE decode
+        self.load_models_to_device(['vae'])
+        latents = inputs_shared["latents"] / self.vae.scaling_factor
+        image = self.vae.decode(latents)
+        image = self.vae_output_to_image(image)
+        self.load_models_to_device([])
+
+        return image
+
+
+class SDUnit_ShapeChecker(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width"),
+            output_params=("height", "width"),
+        )
+
+    def process(self, pipe: StableDiffusionPipeline, height, width):
+        height, width = pipe.check_resize_height_width(height, width)
+        return {"height": height, "width": width}
+
+
+class SDUnit_PromptEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            seperate_cfg=True,
+            input_params_posi={"prompt": "prompt"},
+            input_params_nega={"prompt": "negative_prompt"},
+            output_params=("prompt_embeds",),
+            onload_model_names=("text_encoder",)
+        )
+
+    def encode_prompt(
+        self,
+        pipe: StableDiffusionPipeline,
+        prompt: str,
+        device: torch.device,
+    ) -> torch.Tensor:
+        text_inputs = pipe.tokenizer(
+            prompt,
+            padding="max_length",
+            max_length=pipe.tokenizer.model_max_length,
+            truncation=True,
+            return_tensors="pt",
+        )
+        text_input_ids = text_inputs.input_ids.to(device)
+        prompt_embeds = pipe.text_encoder(text_input_ids)
+        # TextEncoder returns (last_hidden_state, hidden_states) or just last_hidden_state.
+        # last_hidden_state is the post-final-layer-norm output, matching diffusers encode_prompt.
+        if isinstance(prompt_embeds, tuple):
+            prompt_embeds = prompt_embeds[0]
+        return prompt_embeds
+
+    def process(self, pipe: StableDiffusionPipeline, prompt):
+        pipe.load_models_to_device(self.onload_model_names)
+        prompt_embeds = self.encode_prompt(pipe, prompt, pipe.device)
+        return {"prompt_embeds": prompt_embeds}
+
+
+class SDUnit_NoiseInitializer(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width", "seed", "rand_device"),
+            output_params=("noise",),
+        )
+
+    def process(self, pipe: StableDiffusionPipeline, height, width, seed, rand_device):
+        noise = pipe.generate_noise(
+            (1, pipe.unet.in_channels, height // 8, width // 8),
+            seed=seed, rand_device=rand_device, rand_torch_dtype=pipe.torch_dtype
+        )
+        return {"noise": noise}
+
+
+class SDUnit_InputImageEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("input_image", "noise"),
+            output_params=("latents", "input_latents"),
+            onload_model_names=("vae",),
+        )
+
+    def process(self, pipe: StableDiffusionPipeline, input_image, noise):
+        if input_image is None:
+            return {"latents": noise}
+        pipe.load_models_to_device(self.onload_model_names)
+        input_tensor = pipe.preprocess_image(input_image)
+        input_latents = pipe.vae.encode(input_tensor).sample() * pipe.vae.scaling_factor
+        latents = pipe.scheduler.add_noise(input_latents, noise, pipe.scheduler.timesteps[0])
+        if pipe.scheduler.training:
+            return {"latents": latents, "input_latents": input_latents}
+        else:
+            return {"latents": latents}
+
+
+def model_fn_stable_diffusion(
+    unet: UNet2DConditionModel,
+    latents=None,
+    timestep=None,
+    prompt_embeds=None,
+    cross_attention_kwargs=None,
+    timestep_cond=None,
+    added_cond_kwargs=None,
+    **kwargs,
+):
+    # SD timestep is already in 0-999 range, no scaling needed
+    noise_pred = unet(
+        latents,
+        timestep,
+        encoder_hidden_states=prompt_embeds,
+        cross_attention_kwargs=cross_attention_kwargs,
+        timestep_cond=timestep_cond,
+        added_cond_kwargs=added_cond_kwargs,
+        return_dict=False,
+    )[0]
+    return noise_pred

diffsynth/pipelines/stable_diffusion_xl.py

@@ -0,0 +1,331 @@
+import torch
+from PIL import Image
+from tqdm import tqdm
+from typing import Union
+
+from ..core.device.npu_compatible_device import get_device_type
+from ..diffusion.ddim_scheduler import DDIMScheduler
+from ..core import ModelConfig
+from ..diffusion.base_pipeline import BasePipeline, PipelineUnit
+
+from transformers import AutoTokenizer, CLIPTextModel
+from ..models.stable_diffusion_text_encoder import SDTextEncoder
+from ..models.stable_diffusion_xl_unet import SDXLUNet2DConditionModel
+from ..models.stable_diffusion_xl_text_encoder import SDXLTextEncoder2
+from ..models.stable_diffusion_vae import StableDiffusionVAE
+
+
+def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
+    """Rescale noise_cfg based on guidance_rescale to prevent overexposure.
+
+    Based on Section 3.4 from "Common Diffusion Noise Schedules and Sample Steps are Flawed"
+    https://huggingface.co/papers/2305.08891
+    """
+    std_text = noise_pred_text.std(dim=list(range(1, noise_pred_text.ndim)), keepdim=True)
+    std_cfg = noise_cfg.std(dim=list(range(1, noise_cfg.ndim)), keepdim=True)
+    noise_pred_rescaled = noise_cfg * (std_text / std_cfg)
+    noise_cfg = guidance_rescale * noise_pred_rescaled + (1 - guidance_rescale) * noise_cfg
+    return noise_cfg
+
+
+class StableDiffusionXLPipeline(BasePipeline):
+
+    def __init__(self, device=get_device_type(), torch_dtype=torch.bfloat16):
+        super().__init__(
+            device=device, torch_dtype=torch_dtype,
+            height_division_factor=8, width_division_factor=8,
+        )
+        self.scheduler = DDIMScheduler()
+        self.text_encoder: SDTextEncoder = None
+        self.text_encoder_2: SDXLTextEncoder2 = None
+        self.unet: SDXLUNet2DConditionModel = None
+        self.vae: StableDiffusionVAE = None
+        self.tokenizer: AutoTokenizer = None
+        self.tokenizer_2: AutoTokenizer = None
+
+        self.in_iteration_models = ("unet",)
+        self.units = [
+            SDXLUnit_ShapeChecker(),
+            SDXLUnit_PromptEmbedder(),
+            SDXLUnit_NoiseInitializer(),
+            SDXLUnit_InputImageEmbedder(),
+            SDXLUnit_AddTimeIdsComputer(),
+        ]
+        self.model_fn = model_fn_stable_diffusion_xl
+        self.compilable_models = ["unet"]
+
+    @staticmethod
+    def from_pretrained(
+        torch_dtype: torch.dtype = torch.bfloat16,
+        device: Union[str, torch.device] = get_device_type(),
+        model_configs: list[ModelConfig] = [],
+        tokenizer_config: ModelConfig = None,
+        tokenizer_2_config: ModelConfig = None,
+        vram_limit: float = None,
+    ):
+        pipe = StableDiffusionXLPipeline(device=device, torch_dtype=torch_dtype)
+        # Override vram_config to use the specified torch_dtype for all models
+        for mc in model_configs:
+            mc._vram_config_override = {
+                'onload_dtype': torch_dtype,
+                'computation_dtype': torch_dtype,
+            }
+        model_pool = pipe.download_and_load_models(model_configs, vram_limit)
+        pipe.text_encoder = model_pool.fetch_model("stable_diffusion_text_encoder")
+        pipe.text_encoder_2 = model_pool.fetch_model("stable_diffusion_xl_text_encoder")
+        pipe.unet = model_pool.fetch_model("stable_diffusion_xl_unet")
+        pipe.vae = model_pool.fetch_model("stable_diffusion_xl_vae")
+        if tokenizer_config is not None:
+            tokenizer_config.download_if_necessary()
+            pipe.tokenizer = AutoTokenizer.from_pretrained(tokenizer_config.path)
+        if tokenizer_2_config is not None:
+            tokenizer_2_config.download_if_necessary()
+            pipe.tokenizer_2 = AutoTokenizer.from_pretrained(tokenizer_2_config.path)
+        pipe.vram_management_enabled = pipe.check_vram_management_state()
+        return pipe
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        prompt: str,
+        negative_prompt: str = "",
+        cfg_scale: float = 5.0,
+        height: int = 1024,
+        width: int = 1024,
+        seed: int = None,
+        rand_device: str = "cpu",
+        num_inference_steps: int = 50,
+        guidance_rescale: float = 0.0,
+        progress_bar_cmd=tqdm,
+    ):
+        # 1. Scheduler
+        self.scheduler.set_timesteps(num_inference_steps)
+
+        # 2. Three-dict input preparation
+        inputs_posi = {
+            "prompt": prompt,
+        }
+        inputs_nega = {
+            "prompt": negative_prompt,
+        }
+        inputs_shared = {
+            "cfg_scale": cfg_scale,
+            "height": height, "width": width,
+            "seed": seed, "rand_device": rand_device,
+            "guidance_rescale": guidance_rescale,
+            "crops_coords_top_left": (0, 0),
+        }
+
+        # 3. Unit chain execution
+        for unit in self.units:
+            inputs_shared, inputs_posi, inputs_nega = self.unit_runner(
+                unit, self, inputs_shared, inputs_posi, inputs_nega
+            )
+
+        # 4. Denoise loop
+        self.load_models_to_device(self.in_iteration_models)
+        models = {name: getattr(self, name) for name in self.in_iteration_models}
+        for progress_id, timestep in enumerate(progress_bar_cmd(self.scheduler.timesteps)):
+            timestep = timestep.unsqueeze(0).to(dtype=self.torch_dtype, device=self.device)
+            noise_pred = self.cfg_guided_model_fn(
+                self.model_fn, cfg_scale,
+                inputs_shared, inputs_posi, inputs_nega,
+                **models, timestep=timestep, progress_id=progress_id
+            )
+
+            # Apply guidance_rescale
+            if guidance_rescale > 0.0:
+                # cfg_guided_model_fn already applied CFG, now apply rescale
+                # We need the text-only prediction for rescale
+                noise_pred_text = self.model_fn(
+                    self.unet,
+                    inputs_shared["latents"],
+                    timestep,
+                    inputs_posi["prompt_embeds"],
+                    pooled_prompt_embeds=inputs_posi["pooled_prompt_embeds"],
+                    add_time_ids=inputs_posi["add_time_ids"],
+                )
+                noise_pred = rescale_noise_cfg(
+                    noise_pred, noise_pred_text, guidance_rescale=guidance_rescale
+                )
+
+            inputs_shared["latents"] = self.step(
+                self.scheduler, progress_id=progress_id, noise_pred=noise_pred, **inputs_shared
+            )
+
+        # 6. VAE decode
+        self.load_models_to_device(['vae'])
+        latents = inputs_shared["latents"] / self.vae.scaling_factor
+        image = self.vae.decode(latents)
+        image = self.vae_output_to_image(image)
+        self.load_models_to_device([])
+
+        return image
+
+
+class SDXLUnit_ShapeChecker(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width"),
+            output_params=("height", "width"),
+        )
+
+    def process(self, pipe: StableDiffusionXLPipeline, height, width):
+        height, width = pipe.check_resize_height_width(height, width)
+        return {"height": height, "width": width}
+
+
+class SDXLUnit_PromptEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            seperate_cfg=True,
+            input_params_posi={"prompt": "prompt"},
+            input_params_nega={"prompt": "prompt"},
+            output_params=("prompt_embeds", "pooled_prompt_embeds"),
+            onload_model_names=("text_encoder", "text_encoder_2")
+        )
+
+    def encode_prompt(
+        self,
+        pipe: StableDiffusionXLPipeline,
+        prompt: str,
+        device: torch.device,
+    ) -> tuple:
+        """Encode prompt using both text encoders (same prompt for both).
+
+        Returns (prompt_embeds, pooled_prompt_embeds):
+          - prompt_embeds: concat(encoder1_output, encoder2_output) -> (B, 77, 2048)
+          - pooled_prompt_embeds: encoder2 pooled output -> (B, 1280)
+        """
+        # Text Encoder 1 (CLIP-L, 768-dim)
+        text_input_ids_1 = pipe.tokenizer(
+            prompt,
+            padding="max_length",
+            max_length=pipe.tokenizer.model_max_length,
+            truncation=True,
+            return_tensors="pt",
+        ).input_ids.to(device)
+        prompt_embeds_1 = pipe.text_encoder(text_input_ids_1)
+        if isinstance(prompt_embeds_1, tuple):
+            prompt_embeds_1 = prompt_embeds_1[0]
+
+        # Text Encoder 2 (CLIP-bigG, 1280-dim) — uses penultimate hidden states + pooled
+        text_input_ids_2 = pipe.tokenizer_2(
+            prompt,
+            padding="max_length",
+            max_length=pipe.tokenizer_2.model_max_length,
+            truncation=True,
+            return_tensors="pt",
+        ).input_ids.to(device)
+        # SDXLTextEncoder2 forward returns (text_embeds/pooled, hidden_states_tuple)
+        pooled_prompt_embeds, hidden_states = pipe.text_encoder_2(text_input_ids_2, output_hidden_states=True)
+        # Use penultimate hidden state (same as diffusers: hidden_states[-2])
+        prompt_embeds_2 = hidden_states[-2]
+
+        # Concatenate both encoder outputs along feature dimension
+        prompt_embeds = torch.cat([prompt_embeds_1, prompt_embeds_2], dim=-1)
+
+        return prompt_embeds, pooled_prompt_embeds
+
+    def process(self, pipe: StableDiffusionXLPipeline, prompt):
+        pipe.load_models_to_device(self.onload_model_names)
+        prompt_embeds, pooled_prompt_embeds = self.encode_prompt(pipe, prompt, pipe.device)
+        return {"prompt_embeds": prompt_embeds, "pooled_prompt_embeds": pooled_prompt_embeds}
+
+
+class SDXLUnit_NoiseInitializer(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width", "seed", "rand_device"),
+            output_params=("noise",),
+        )
+
+    def process(self, pipe: StableDiffusionXLPipeline, height, width, seed, rand_device):
+        noise = pipe.generate_noise(
+            (1, pipe.unet.in_channels, height // 8, width // 8),
+            seed=seed, rand_device=rand_device, rand_torch_dtype=pipe.torch_dtype
+        )
+        return {"noise": noise}
+
+
+class SDXLUnit_InputImageEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("input_image", "noise"),
+            output_params=("latents", "input_latents"),
+            onload_model_names=("vae",),
+        )
+
+    def process(self, pipe: StableDiffusionXLPipeline, input_image, noise):
+        if input_image is None:
+            return {"latents": noise}
+        pipe.load_models_to_device(self.onload_model_names)
+        input_tensor = pipe.preprocess_image(input_image)
+        input_latents = pipe.vae.encode(input_tensor).sample() * pipe.vae.scaling_factor
+        latents = pipe.scheduler.add_noise(input_latents, noise, pipe.scheduler.timesteps[0])
+        if pipe.scheduler.training:
+            return {"latents": latents, "input_latents": input_latents}
+        else:
+            return {"latents": latents}
+
+
+class SDXLUnit_AddTimeIdsComputer(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("height", "width"),
+            output_params=("add_time_ids",),
+        )
+
+    def _get_add_time_ids(self, pipe, original_size, crops_coords_top_left, target_size, dtype, text_encoder_projection_dim):
+        add_time_ids = list(original_size + crops_coords_top_left + target_size)
+        expected_add_embed_dim = pipe.unet.add_embedding.linear_1.in_features
+        addition_time_embed_dim = pipe.unet.add_time_proj.num_channels
+        passed_add_embed_dim = addition_time_embed_dim * len(add_time_ids) + text_encoder_projection_dim
+        if expected_add_embed_dim != passed_add_embed_dim:
+            raise ValueError(
+                f"Model expects an added time embedding vector of length {expected_add_embed_dim}, "
+                f"but a vector of {passed_add_embed_dim} was created."
+            )
+        add_time_ids = torch.tensor([add_time_ids], dtype=dtype, device=pipe.device)
+        return add_time_ids
+
+    def process(self, pipe: StableDiffusionXLPipeline, height, width):
+        original_size = (height, width)
+        target_size = (height, width)
+        crops_coords_top_left = (0, 0)
+
+        text_encoder_projection_dim = pipe.text_encoder_2.config.projection_dim
+        add_time_ids = self._get_add_time_ids(
+            pipe, original_size, crops_coords_top_left, target_size,
+            dtype=pipe.torch_dtype,
+            text_encoder_projection_dim=text_encoder_projection_dim,
+        )
+        return {"add_time_ids": add_time_ids}
+
+
+def model_fn_stable_diffusion_xl(
+    unet: SDXLUNet2DConditionModel,
+    latents=None,
+    timestep=None,
+    prompt_embeds=None,
+    pooled_prompt_embeds=None,
+    add_time_ids=None,
+    cross_attention_kwargs=None,
+    timestep_cond=None,
+    **kwargs,
+):
+    """SDXL model forward with added_cond_kwargs for micro-conditioning."""
+    added_cond_kwargs = {
+        "text_embeds": pooled_prompt_embeds,
+        "time_ids": add_time_ids,
+    }
+    noise_pred = unet(
+        latents,
+        timestep,
+        encoder_hidden_states=prompt_embeds,
+        added_cond_kwargs=added_cond_kwargs,
+        cross_attention_kwargs=cross_attention_kwargs,
+        timestep_cond=timestep_cond,
+        return_dict=False,
+    )[0]
+    return noise_pred

diffsynth/pipelines/wan_video.py

@@ -75,15 +75,19 @@ class WanVideoPipeline(BasePipeline):
             WanVideoUnit_TeaCache(),
             WanVideoUnit_CfgMerger(),
             WanVideoUnit_LongCatVideo(),
+            WanVideoUnit_WanToDance_ProcessInputs(),
+            WanVideoUnit_WanToDance_RefImageEmbedder(),
+            WanVideoUnit_WanToDance_ImageKeyframesEmbedder(),
         ]
         self.post_units = [
             WanVideoPostUnit_S2V(),
         ]
         self.model_fn = model_fn_wan_video
+        self.compilable_models = ["dit", "dit2"]
 
 
     def enable_usp(self):
-        from ..utils.xfuser import get_sequence_parallel_world_size, usp_attn_forward, usp_dit_forward
+        from ..utils.xfuser import get_sequence_parallel_world_size, usp_attn_forward, usp_dit_forward, usp_vace_forward
 
         for block in self.dit.blocks:
             block.self_attn.forward = types.MethodType(usp_attn_forward, block.self_attn)
@@ -92,6 +96,14 @@ class WanVideoPipeline(BasePipeline):
             for block in self.dit2.blocks:
                 block.self_attn.forward = types.MethodType(usp_attn_forward, block.self_attn)
             self.dit2.forward = types.MethodType(usp_dit_forward, self.dit2)
+        if self.vace is not None:
+            for block in self.vace.vace_blocks:
+                block.self_attn.forward = types.MethodType(usp_attn_forward, block.self_attn)
+            self.vace.forward = types.MethodType(usp_vace_forward, self.vace)
+        if self.vace2 is not None:
+            for block in self.vace2.vace_blocks:
+                block.self_attn.forward = types.MethodType(usp_attn_forward, block.self_attn)
+            self.vace2.forward = types.MethodType(usp_vace_forward, self.vace2)
         self.sp_size = get_sequence_parallel_world_size()
         self.use_unified_sequence_parallel = True
 
@@ -244,6 +256,13 @@ class WanVideoPipeline(BasePipeline):
         # Teacache
         tea_cache_l1_thresh: Optional[float] = None,
         tea_cache_model_id: Optional[str] = "",
+        # WanToDance
+        wantodance_music_path: Optional[str] = None,
+        wantodance_reference_image: Optional[Image.Image] = None,
+        wantodance_fps: Optional[float] = 30,
+        wantodance_keyframes: Optional[list[Image.Image]] = None,
+        wantodance_keyframes_mask: Optional[list[int]] = None,
+        framewise_decoding: bool = False,
         # progress_bar
         progress_bar_cmd=tqdm,
         output_type: Optional[Literal["quantized", "floatpoint"]] = "quantized",
@@ -280,6 +299,9 @@ class WanVideoPipeline(BasePipeline):
             "input_audio": input_audio, "audio_sample_rate": audio_sample_rate, "s2v_pose_video": s2v_pose_video, "audio_embeds": audio_embeds, "s2v_pose_latents": s2v_pose_latents, "motion_video": motion_video,
             "animate_pose_video": animate_pose_video, "animate_face_video": animate_face_video, "animate_inpaint_video": animate_inpaint_video, "animate_mask_video": animate_mask_video,
             "vap_video": vap_video, 
+            "wantodance_music_path": wantodance_music_path, "wantodance_reference_image": wantodance_reference_image, "wantodance_fps": wantodance_fps,
+            "wantodance_keyframes": wantodance_keyframes, "wantodance_keyframes_mask": wantodance_keyframes_mask,
+            "framewise_decoding": framewise_decoding,
         }
         for unit in self.units:
             inputs_shared, inputs_posi, inputs_nega = self.unit_runner(unit, self, inputs_shared, inputs_posi, inputs_nega)
@@ -325,7 +347,10 @@ class WanVideoPipeline(BasePipeline):
             inputs_shared, _, _ = self.unit_runner(unit, self, inputs_shared, inputs_posi, inputs_nega)
         # Decode
         self.load_models_to_device(['vae'])
-        video = self.vae.decode(inputs_shared["latents"], device=self.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride)
+        if framewise_decoding:
+            video = self.vae.decode_framewise(inputs_shared["latents"], device=self.device)
+        else:
+            video = self.vae.decode(inputs_shared["latents"], device=self.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride)
         if output_type == "quantized":
             video = self.vae_output_to_video(video)
         elif output_type == "floatpoint":
@@ -371,17 +396,20 @@ class WanVideoUnit_NoiseInitializer(PipelineUnit):
 class WanVideoUnit_InputVideoEmbedder(PipelineUnit):
     def __init__(self):
         super().__init__(
-            input_params=("input_video", "noise", "tiled", "tile_size", "tile_stride", "vace_reference_image"),
+            input_params=("input_video", "noise", "tiled", "tile_size", "tile_stride", "vace_reference_image", "framewise_decoding"),
             output_params=("latents", "input_latents"),
             onload_model_names=("vae",)
         )
 
-    def process(self, pipe: WanVideoPipeline, input_video, noise, tiled, tile_size, tile_stride, vace_reference_image):
+    def process(self, pipe: WanVideoPipeline, input_video, noise, tiled, tile_size, tile_stride, vace_reference_image, framewise_decoding):
         if input_video is None:
             return {"latents": noise}
         pipe.load_models_to_device(self.onload_model_names)
         input_video = pipe.preprocess_video(input_video)
-        input_latents = pipe.vae.encode(input_video, device=pipe.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride).to(dtype=pipe.torch_dtype, device=pipe.device)
+        if framewise_decoding:
+            input_latents = pipe.vae.encode_framewise(input_video, device=pipe.device)
+        else:
+            input_latents = pipe.vae.encode(input_video, device=pipe.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride).to(dtype=pipe.torch_dtype, device=pipe.device)
         if vace_reference_image is not None:
             if not isinstance(vace_reference_image, list):
                 vace_reference_image = [vace_reference_image]
@@ -1018,6 +1046,111 @@ class WanVideoUnit_LongCatVideo(PipelineUnit):
         return {"longcat_latents": longcat_latents}
 
 
+class WanVideoUnit_WanToDance_ProcessInputs(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            take_over=True,
+        )
+
+    def get_music_base_feature(self, music_path, fps=30):
+        import librosa
+        hop_length = 512
+        sr = fps * hop_length
+        data, sr = librosa.load(music_path, sr=sr)
+        sr = 22050 
+        envelope = librosa.onset.onset_strength(y=data, sr=sr)
+        mfcc = librosa.feature.mfcc(y=data, sr=sr, n_mfcc=20).T  
+        chroma = librosa.feature.chroma_cens(
+            y=data, sr=sr, hop_length=hop_length, n_chroma=12
+        ).T 
+        peak_idxs = librosa.onset.onset_detect(
+            onset_envelope=envelope.flatten(), sr=sr, hop_length=hop_length
+        )
+        peak_onehot = np.zeros_like(envelope, dtype=np.float32)
+        peak_onehot[peak_idxs] = 1.0
+        start_bpm = librosa.beat.tempo(y=librosa.load(music_path)[0])[0]
+        _, beat_idxs = librosa.beat.beat_track(
+            onset_envelope=envelope,
+            sr=sr,
+            hop_length=hop_length,
+            start_bpm=start_bpm,
+            tightness=100,
+        )
+        beat_onehot = np.zeros_like(envelope, dtype=np.float32)
+        beat_onehot[beat_idxs] = 1.0  
+        audio_feature = np.concatenate(
+            [envelope[:, None], mfcc, chroma, peak_onehot[:, None], beat_onehot[:, None]],
+            axis=-1,
+        )
+        return torch.from_numpy(audio_feature)
+
+    def process(self, pipe: WanVideoPipeline, inputs_shared, inputs_posi, inputs_nega):
+        if pipe.dit.wantodance_enable_global:
+            inputs_nega["skip_9th_layer"] = True
+        if inputs_shared.get("wantodance_music_path", None) is not None:
+            inputs_shared["music_feature"] = self.get_music_base_feature(inputs_shared["wantodance_music_path"]).to(dtype=pipe.torch_dtype, device=pipe.device)
+        return inputs_shared, inputs_posi, inputs_nega
+
+
+class WanVideoUnit_WanToDance_RefImageEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("wantodance_reference_image", "num_frames", "height", "width", "tiled", "tile_size", "tile_stride"),
+            output_params=("wantodance_refimage_feature",),
+            onload_model_names=("image_encoder", "vae")
+        )
+
+    def process(self, pipe: WanVideoPipeline, wantodance_reference_image, num_frames, height, width, tiled, tile_size, tile_stride):
+        if wantodance_reference_image is None:
+            return {}
+        pipe.load_models_to_device(self.onload_model_names)
+        if isinstance(wantodance_reference_image, list):
+            wantodance_reference_image = wantodance_reference_image[0]
+        image = pipe.preprocess_image(wantodance_reference_image.resize((width, height))).to(pipe.device) # B,C,H,W;B=1
+        refimage_feature = pipe.image_encoder.encode_image([image])
+        refimage_feature = refimage_feature.to(dtype=pipe.torch_dtype, device=pipe.device)
+        return {"wantodance_refimage_feature": refimage_feature}
+
+
+class WanVideoUnit_WanToDance_ImageKeyframesEmbedder(PipelineUnit):
+    def __init__(self):
+        super().__init__(
+            input_params=("wantodance_keyframes", "wantodance_keyframes_mask", "num_frames", "height", "width", "tiled", "tile_size", "tile_stride"),
+            output_params=("clip_feature", "y"),
+            onload_model_names=("image_encoder", "vae")
+        )
+
+    def process(self, pipe: WanVideoPipeline, wantodance_keyframes, wantodance_keyframes_mask, num_frames, height, width, tiled, tile_size, tile_stride):
+        if wantodance_keyframes is None:
+            return {}
+        wantodance_keyframes_mask = torch.tensor(wantodance_keyframes_mask)
+        pipe.load_models_to_device(self.onload_model_names)
+        images = []
+        for input_image in wantodance_keyframes:
+            input_image = pipe.preprocess_image(input_image.resize((width, height))).to(pipe.device)
+            images.append(input_image)
+    
+        clip_context = pipe.image_encoder.encode_image(images[:1]) # 取第一帧作为clip输入
+        msk = torch.zeros(1, num_frames, height//8, width//8, device=pipe.device)
+        msk[:, wantodance_keyframes_mask==1, :, :] = torch.ones(1, height//8, width//8, device=pipe.device) # set keyframes mask to 1
+        
+        images = [image.transpose(0, 1) for image in images]  # 3, num_frames, h, w
+        images = torch.concat(images, dim=1) 
+        vae_input = images
+
+        msk = torch.concat([torch.repeat_interleave(msk[:, 0:1], repeats=4, dim=1), msk[:, 1:]], dim=1) # expand first frame mask, N to N + 3
+        msk = msk.view(1, msk.shape[1] // 4, 4, height//8, width//8)
+        msk = msk.transpose(1, 2)[0]
+        
+        y = pipe.vae.encode([vae_input.to(dtype=pipe.torch_dtype, device=pipe.device)], device=pipe.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride)[0]
+        y = y.to(dtype=pipe.torch_dtype, device=pipe.device)
+        y = torch.concat([msk, y])
+        y = y.unsqueeze(0)
+        clip_context = clip_context.to(dtype=pipe.torch_dtype, device=pipe.device)
+        y = y.to(dtype=pipe.torch_dtype, device=pipe.device)
+        return {"clip_feature": clip_context, "y": y}
+
+
 class TeaCache:
     def __init__(self, num_inference_steps, rel_l1_thresh, model_id):
         self.num_inference_steps = num_inference_steps
@@ -1123,6 +1256,22 @@ class TemporalTiler_BCTHW:
         return value
 
 
+def wantodance_get_single_freqs(freqs, frame_num, fps):
+    total_frame = int(30.0 / (fps + 1e-6) * frame_num + 0.5)
+    interval_frame = 30.0 / (fps + 1e-6)
+    freqs_0 = freqs[:total_frame]
+    freqs_new = torch.zeros((frame_num, freqs_0.shape[1]), device=freqs_0.device, dtype=freqs_0.dtype)
+    freqs_new[0] = freqs_0[0]
+    freqs_new[-1] = freqs_0[total_frame - 1]
+    for i in range(1, frame_num-1):
+        pos = i * interval_frame
+        low_idx = int(pos)
+        high_idx = min(low_idx + 1, total_frame - 1)
+        weight_high = pos - low_idx
+        weight_low = 1.0 - weight_high
+        freqs_new[i] = freqs_0[low_idx] * weight_low + freqs_0[high_idx] * weight_high
+    return freqs_new
+
 
 def model_fn_wan_video(
     dit: WanModel,
@@ -1158,6 +1307,10 @@ def model_fn_wan_video(
     use_gradient_checkpointing_offload: bool = False,
     control_camera_latents_input = None,
     fuse_vae_embedding_in_latents: bool = False,
+    wantodance_refimage_feature = None,
+    wantodance_fps: float = 30.0,
+    music_feature = None,
+    skip_9th_layer: bool = False,
     **kwargs,
 ):
     if sliding_window_size is not None and sliding_window_stride is not None:
@@ -1255,7 +1408,10 @@ def model_fn_wan_video(
         context = torch.cat([clip_embdding, context], dim=1)
         
     # Camera control
-    x = dit.patchify(x, control_camera_latents_input)
+    if hasattr(dit, "wantodance_enable_global") and dit.wantodance_enable_global and int(wantodance_fps + 0.5) != 30:
+        x = dit.patchify(x, control_camera_latents_input, enable_wantodance_global=True)
+    else:
+        x = dit.patchify(x, control_camera_latents_input)
     
     # Animate
     if pose_latents is not None and face_pixel_values is not None:
@@ -1303,14 +1459,61 @@ def model_fn_wan_video(
         tea_cache_update = tea_cache.check(dit, x, t_mod)
     else:
         tea_cache_update = False
-        
-    if vace_context is not None:
-        vace_hints = vace(
-            x, vace_context, context, t_mod, freqs,
-            use_gradient_checkpointing=use_gradient_checkpointing,
-            use_gradient_checkpointing_offload=use_gradient_checkpointing_offload
-        )
-    
+
+    # WanToDance
+    if hasattr(dit, "wantodance_enable_global") and dit.wantodance_enable_global:
+        if wantodance_refimage_feature is not None:
+            refimage_feature_embedding = dit.img_emb_refimage(wantodance_refimage_feature)
+            context = torch.cat([refimage_feature_embedding, context], dim=1)
+        if (dit.wantodance_enable_dynamicfps or dit.wantodance_enable_unimodel) and int(wantodance_fps + 0.5) != 30: 
+            freqs_0 = wantodance_get_single_freqs(dit.freqs[0], f, wantodance_fps)
+            freqs = torch.cat([
+                freqs_0.view(f, 1, 1, -1).expand(f, h, w, -1),
+                dit.freqs[1][:h].view(1, h, 1, -1).expand(f, h, w, -1),
+                dit.freqs[2][:w].view(1, 1, w, -1).expand(f, h, w, -1)
+            ], dim=-1).reshape(f * h * w, 1, -1).to(x.device)
+        if dit.wantodance_enable_global or dit.wantodance_enable_dynamicfps or dit.wantodance_enable_unimodel:
+            if use_unified_sequence_parallel:
+                length = int(float(music_feature.shape[0]) / get_sequence_parallel_world_size()) * get_sequence_parallel_world_size()
+                music_feature = music_feature[:length]
+                music_feature = torch.chunk(music_feature, get_sequence_parallel_world_size(), dim=0)[get_sequence_parallel_rank()]
+            if not dit.training:
+                dit.music_encoder.to(x.device, dtype=x.dtype) # only evaluation
+            music_feature = music_feature.to(x.device, dtype=x.dtype)
+            music_feature = dit.music_projection(music_feature)
+            music_feature = dit.music_encoder(music_feature)
+            if music_feature.dim() == 2:
+                music_feature = music_feature.unsqueeze(0)
+            if use_unified_sequence_parallel:
+                if dist.is_initialized() and dist.get_world_size() > 1:
+                    music_feature = get_sp_group().all_gather(music_feature, dim=1)
+            music_feature = music_feature.unsqueeze(1) # [1, 1, 149, 4800]
+            N = 149 
+            M = 4800 
+            music_feature = torch.nn.functional.interpolate(music_feature, size=(N, M), mode='bilinear') 
+            music_feature = music_feature.squeeze(1) # shape: [1, 149, 4800]
+        if music_feature is not None:
+            if music_feature.dim() == 2:
+                music_feature = music_feature.unsqueeze(0)
+            music_feature = music_feature.to(x.device, dtype=x.dtype)
+            interp_mode = 'bilinear'
+            if interp_mode == 'bilinear':
+                frame_num = latents.shape[2] if len(latents.shape) == 5 else latents.shape[1] # 21
+                context_shape_end = context.shape[2] ## 14B 5120
+                music_feature = music_feature.unsqueeze(1) # shape: [1, 1, 149, 4800]
+                if use_unified_sequence_parallel:
+                    N = int(float(frame_num * 8) / get_sequence_parallel_world_size()) * get_sequence_parallel_world_size()
+                else:
+                    N = frame_num * 8
+                music_feature = torch.nn.functional.interpolate(music_feature, size=(N, context_shape_end), mode='bilinear') 
+                music_feature = music_feature.squeeze(1) # shape: [1, N, context_shape_end]
+                if use_unified_sequence_parallel:
+                    dit.merged_audio_emb = torch.chunk(music_feature, get_sequence_parallel_world_size(), dim=1)[get_sequence_parallel_rank()]
+                else:
+                    dit.merged_audio_emb = music_feature
+            else: 
+                dit.merged_audio_emb = music_feature
+
     # blocks
     if use_unified_sequence_parallel:
         if dist.is_initialized() and dist.get_world_size() > 1:
@@ -1318,6 +1521,13 @@ def model_fn_wan_video(
             pad_shape = chunks[0].shape[1] - chunks[-1].shape[1]
             chunks = [torch.nn.functional.pad(chunk, (0, 0, 0, chunks[0].shape[1]-chunk.shape[1]), value=0) for chunk in chunks]
             x = chunks[get_sequence_parallel_rank()]
+
+    if vace_context is not None:
+        vace_hints = vace(
+            x, vace_context, context, t_mod, freqs,
+            use_gradient_checkpointing=use_gradient_checkpointing,
+            use_gradient_checkpointing_offload=use_gradient_checkpointing_offload
+        )
     if tea_cache_update:
         x = tea_cache.update(x)
     else:
@@ -1326,8 +1536,12 @@ def model_fn_wan_video(
                 return vap(block, *inputs)
             return custom_forward
         
+        # Block
         for block_id, block in enumerate(dit.blocks):
-            # Block
+            if skip_9th_layer:
+                # This is only used in WanToDance
+                if block_id == 9:
+                    continue
             if vap is not None and block_id in vap.mot_layers_mapping:
                 if use_gradient_checkpointing_offload:
                     with torch.autograd.graph.save_on_cpu():
@@ -1356,18 +1570,23 @@ def model_fn_wan_video(
             # VACE
             if vace_context is not None and block_id in vace.vace_layers_mapping:
                 current_vace_hint = vace_hints[vace.vace_layers_mapping[block_id]]
-                if use_unified_sequence_parallel and dist.is_initialized() and dist.get_world_size() > 1:
-                    current_vace_hint = torch.chunk(current_vace_hint, get_sequence_parallel_world_size(), dim=1)[get_sequence_parallel_rank()]
-                    current_vace_hint = torch.nn.functional.pad(current_vace_hint, (0, 0, 0, chunks[0].shape[1] - current_vace_hint.shape[1]), value=0)
                 x = x + current_vace_hint * vace_scale
             
             # Animate
             if pose_latents is not None and face_pixel_values is not None:
                 x = animate_adapter.after_transformer_block(block_id, x, motion_vec)
+            
+            # WanToDance
+            if hasattr(dit, "wantodance_enable_music_inject") and dit.wantodance_enable_music_inject:
+                x = dit.wantodance_after_transformer_block(block_id, x)
         if tea_cache is not None:
             tea_cache.store(x)
             
-    x = dit.head(x, t)
+    if hasattr(dit, "wantodance_enable_unimodel") and dit.wantodance_enable_unimodel and int(wantodance_fps + 0.5) != 30:
+        x = dit.head_global(x, t)
+    else:
+        x = dit.head(x, t)
+    
     if use_unified_sequence_parallel:
         if dist.is_initialized() and dist.get_world_size() > 1:
             x = get_sp_group().all_gather(x, dim=1)

diffsynth/pipelines/z_image.py

@@ -54,6 +54,7 @@ class ZImagePipeline(BasePipeline):
             ZImageUnit_PAIControlNet(),
         ]
         self.model_fn = model_fn_z_image
+        self.compilable_models = ["dit"]
     
     
     @staticmethod

diffsynth/utils/state_dict_converters/ernie_image_text_encoder.py

@@ -0,0 +1,21 @@
+def ErnieImageTextEncoderStateDictConverter(state_dict):
+    """
+    Maps checkpoint keys from multimodal Mistral3Model format
+    to text-only Ministral3Model format.
+
+    Checkpoint keys (Mistral3Model):
+        language_model.model.layers.0.input_layernorm.weight
+        language_model.model.norm.weight
+
+    Model keys (ErnieImageTextEncoder → self.model = Ministral3Model):
+        model.layers.0.input_layernorm.weight
+        model.norm.weight
+
+    Mapping: language_model. → model.
+    """
+    new_state_dict = {}
+    for key in state_dict:
+        if key.startswith("language_model.model."):
+            new_key = key.replace("language_model.model.", "model.", 1)
+            new_state_dict[new_key] = state_dict[key]
+    return new_state_dict

diffsynth/utils/state_dict_converters/joyai_image_text_encoder.py

@@ -0,0 +1,20 @@
+def JoyAIImageTextEncoderStateDictConverter(state_dict):
+    """Convert HuggingFace Qwen3VL checkpoint keys to DiffSynth wrapper keys.
+
+    Mapping (checkpoint -> wrapper):
+    - lm_head.weight -> model.lm_head.weight
+    - model.language_model.* -> model.model.language_model.*
+    - model.visual.* -> model.model.visual.*
+    """
+    state_dict_ = {}
+    for key in state_dict:
+        if key == "lm_head.weight":
+            new_key = "model.lm_head.weight"
+        elif key.startswith("model.language_model."):
+            new_key = "model.model." + key[len("model."):]
+        elif key.startswith("model.visual."):
+            new_key = "model.model." + key[len("model."):]
+        else:
+            new_key = key
+        state_dict_[new_key] = state_dict[key]
+    return state_dict_

diffsynth/utils/state_dict_converters/stable_diffusion_text_encoder.py

@@ -0,0 +1,7 @@
+def SDTextEncoderStateDictConverter(state_dict):
+    new_state_dict = {}
+    for key in state_dict:
+        if key.startswith("text_model.") and "position_ids" not in key:
+            new_key = "model." + key
+            new_state_dict[new_key] = state_dict[key]
+    return new_state_dict

diffsynth/utils/state_dict_converters/stable_diffusion_vae.py

@@ -0,0 +1,18 @@
+def SDVAEStateDictConverter(state_dict):
+    new_state_dict = {}
+    for key in state_dict:
+        if ".query." in key:
+            new_key = key.replace(".query.", ".to_q.")
+            new_state_dict[new_key] = state_dict[key]
+        elif ".key." in key:
+            new_key = key.replace(".key.", ".to_k.")
+            new_state_dict[new_key] = state_dict[key]
+        elif ".value." in key:
+            new_key = key.replace(".value.", ".to_v.")
+            new_state_dict[new_key] = state_dict[key]
+        elif ".proj_attn." in key:
+            new_key = key.replace(".proj_attn.", ".to_out.0.")
+            new_state_dict[new_key] = state_dict[key]
+        else:
+            new_state_dict[key] = state_dict[key]
+    return new_state_dict

diffsynth/utils/state_dict_converters/stable_diffusion_xl_text_encoder.py

@@ -0,0 +1,13 @@
+import torch
+
+def SDXLTextEncoder2StateDictConverter(state_dict):
+    new_state_dict = {}
+    for key in state_dict:
+        if key == "text_projection.weight":
+            val = state_dict[key]
+            new_state_dict["model.text_projection.weight"] = val.float() if val.dtype == torch.float16 else val
+        elif key.startswith("text_model.") and "position_ids" not in key:
+            new_key = "model." + key
+            val = state_dict[key]
+            new_state_dict[new_key] = val.float() if val.dtype == torch.float16 else val
+    return new_state_dict

diffsynth/utils/state_dict_converters/z_image_dit.py

@@ -0,0 +1,3 @@
+def ZImageDiTStateDictConverter(state_dict):
+    state_dict_ = {name.replace("model.diffusion_model.", ""): state_dict[name] for name in state_dict}
+    return state_dict_

diffsynth/utils/xfuser/__init__.py

@@ -1 +1 @@
-from .xdit_context_parallel import usp_attn_forward, usp_dit_forward, get_sequence_parallel_world_size, initialize_usp, get_current_chunk, gather_all_chunks
+from .xdit_context_parallel import usp_attn_forward, usp_dit_forward, usp_vace_forward, get_sequence_parallel_world_size, initialize_usp, get_current_chunk, gather_all_chunks

diffsynth/utils/xfuser/xdit_context_parallel.py

@@ -117,6 +117,39 @@ def usp_dit_forward(self,
     return x
 
 
+def usp_vace_forward(
+    self, x, vace_context, context, t_mod, freqs,
+    use_gradient_checkpointing: bool = False,
+    use_gradient_checkpointing_offload: bool = False,
+):
+    # Compute full sequence length from the sharded x
+    full_seq_len = x.shape[1] * get_sequence_parallel_world_size()
+
+    # Embed vace_context via patch embedding
+    c = [self.vace_patch_embedding(u.unsqueeze(0)) for u in vace_context]
+    c = [u.flatten(2).transpose(1, 2) for u in c]
+    c = torch.cat([
+        torch.cat([u, u.new_zeros(1, full_seq_len - u.size(1), u.size(2))],
+                  dim=1) for u in c
+    ])
+
+    # Chunk VACE context along sequence dim BEFORE processing through blocks
+    c = torch.chunk(c, get_sequence_parallel_world_size(), dim=1)[get_sequence_parallel_rank()]
+
+    # Process through vace_blocks (self_attn already monkey-patched to usp_attn_forward)
+    for block in self.vace_blocks:
+        c = gradient_checkpoint_forward(
+            block,
+            use_gradient_checkpointing,
+            use_gradient_checkpointing_offload,
+            c, x, context, t_mod, freqs
+        )
+
+    # Hints are already sharded per-rank
+    hints = torch.unbind(c)[:-1]
+    return hints
+
+
 def usp_attn_forward(self, x, freqs):
     q = self.norm_q(self.q(x))
     k = self.norm_k(self.k(x))

docs/en/Model_Details/Anima.md

@@ -133,7 +133,7 @@ Anima models are trained through [`examples/anima/model_training/train.py`](http
 We provide a sample image dataset for testing:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 For training script details, refer to [Model Training](../Pipeline_Usage/Model_Training.md). For advanced training techniques, see [Training Framework Documentation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/).

docs/en/Model_Details/ERNIE-Image.md

@@ -0,0 +1,134 @@
+# ERNIE-Image
+
+ERNIE-Image is a powerful image generation model with 8B parameters developed by Baidu, featuring a compact and efficient architecture with strong instruction-following capability. Based on an 8B DiT backbone, it delivers performance comparable to larger (20B+) models in certain scenarios while maintaining parameter efficiency. It offers reliable performance in instruction understanding and execution, text generation (English/Chinese/Japanese), and overall stability.
+
+## Installation
+
+Before performing model inference and training, please install DiffSynth-Studio first.
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+For more information on installation, please refer to [Setup Dependencies](../Pipeline_Usage/Setup.md).
+
+## Quick Start
+
+Running the following code will load the [PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image) model for inference. VRAM management is enabled, the framework automatically controls parameter loading based on available VRAM, requiring a minimum of 3G VRAM.
+
+```python
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("output.jpg")
+```
+
+## Model Overview
+
+|Model ID|Inference|Low VRAM Inference|Full Training|Full Training Validation|LoRA Training|LoRA Training Validation|
+|-|-|-|-|-|-|-|
+|[PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference/ERNIE-Image.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference_low_vram/ERNIE-Image.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/full/ERNIE-Image.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/validate_full/ERNIE-Image.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/lora/ERNIE-Image.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/validate_lora/ERNIE-Image.py)|
+|[PaddlePaddle/ERNIE-Image-Turbo](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference/ERNIE-Image-Turbo.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference_low_vram/ERNIE-Image-Turbo.py)|—|—|—|—|
+
+## Model Inference
+
+The model is loaded via `ErnieImagePipeline.from_pretrained`, see [Loading Models](../Pipeline_Usage/Model_Inference.md#loading-models) for details.
+
+The input parameters for `ErnieImagePipeline` inference include:
+
+* `prompt`: The prompt describing the content to appear in the image.
+* `negative_prompt`: The negative prompt describing what should not appear in the image, default value is `""`.
+* `cfg_scale`: Classifier-free guidance parameter, default value is 4.0.
+* `height`: Image height, must be a multiple of 16, default value is 1024.
+* `width`: Image width, must be a multiple of 16, default value is 1024.
+* `seed`: Random seed. Default is `None`, meaning completely random.
+* `rand_device`: The computing device for generating random Gaussian noise matrices, default is `"cuda"`. When set to `cuda`, different GPUs will produce different results.
+* `num_inference_steps`: Number of inference steps, default value is 50.
+
+If VRAM is insufficient, please enable [VRAM Management](../Pipeline_Usage/VRAM_management.md). We provide recommended low-VRAM configurations for each model in the "Model Overview" table above.
+
+## Model Training
+
+ERNIE-Image series models are trained uniformly via [`examples/ernie_image/model_training/train.py`](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/train.py). The script parameters include:
+
+* General Training Parameters
+    * Dataset Configuration
+        * `--dataset_base_path`: Root directory of the dataset.
+        * `--dataset_metadata_path`: Path to the dataset metadata file.
+        * `--dataset_repeat`: Number of dataset repeats per epoch.
+        * `--dataset_num_workers`: Number of processes per DataLoader.
+        * `--data_file_keys`: Field names to load from metadata, typically paths to image or video files, separated by `,`.
+    * Model Loading Configuration
+        * `--model_paths`: Paths to load models from, in JSON format.
+        * `--model_id_with_origin_paths`: Model IDs with original paths, e.g., `"PaddlePaddle/ERNIE-Image:transformer/diffusion_pytorch_model*.safetensors"`, separated by commas.
+        * `--extra_inputs`: Additional input parameters required by the model Pipeline, separated by `,`.
+        * `--fp8_models`: Models to load in FP8 format, currently only supported for models whose parameters are not updated by gradients.
+    * Basic Training Configuration
+        * `--learning_rate`: Learning rate.
+        * `--num_epochs`: Number of epochs.
+        * `--trainable_models`: Trainable models, e.g., `dit`, `vae`, `text_encoder`.
+        * `--find_unused_parameters`: Whether unused parameters exist in DDP training.
+        * `--weight_decay`: Weight decay magnitude.
+        * `--task`: Training task, defaults to `sft`.
+    * Output Configuration
+        * `--output_path`: Path to save the model.
+        * `--remove_prefix_in_ckpt`: Remove prefix in the model's state dict.
+        * `--save_steps`: Interval in training steps to save the model.
+    * LoRA Configuration
+        * `--lora_base_model`: Which model to add LoRA to.
+        * `--lora_target_modules`: Which layers to add LoRA to.
+        * `--lora_rank`: Rank of LoRA.
+        * `--lora_checkpoint`: Path to LoRA checkpoint.
+        * `--preset_lora_path`: Path to preset LoRA checkpoint for LoRA differential training.
+        * `--preset_lora_model`: Which model to integrate preset LoRA into, e.g., `dit`.
+    * Gradient Configuration
+        * `--use_gradient_checkpointing`: Whether to enable gradient checkpointing.
+        * `--use_gradient_checkpointing_offload`: Whether to offload gradient checkpointing to CPU memory.
+        * `--gradient_accumulation_steps`: Number of gradient accumulation steps.
+    * Resolution Configuration
+        * `--height`: Height of the image. Leave empty to enable dynamic resolution.
+        * `--width`: Width of the image. Leave empty to enable dynamic resolution.
+        * `--max_pixels`: Maximum pixel area, images larger than this will be scaled down during dynamic resolution.
+* ERNIE-Image Specific Parameters
+    * `--tokenizer_path`: Path to the tokenizer, leave empty to auto-download from remote.
+
+We provide an example image dataset for testing, which can be downloaded with the following command:
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
+```
+
+We provide recommended training scripts for each model, please refer to the table in "Model Overview" above. For guidance on writing model training scripts, see [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, see [Training Framework Overview](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/FLUX.md

@@ -195,7 +195,7 @@ FLUX series models are uniformly trained through [`examples/flux/model_training/
 We have built a sample image dataset for your testing. You can download this dataset with the following command:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 We have written recommended training scripts for each model, please refer to the table in the "Model Overview" section above. For how to write model training scripts, please refer to [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, please refer to [Training Framework Detailed Explanation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/FLUX2.md

@@ -145,7 +145,7 @@ FLUX.2 series models are uniformly trained through [`examples/flux2/model_traini
 We have built a sample image dataset for your testing. You can download this dataset with the following command:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 We have written recommended training scripts for each model, please refer to the table in the "Model Overview" section above. For how to write model training scripts, please refer to [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, please refer to [Training Framework Detailed Explanation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/JoyAI-Image.md

@@ -0,0 +1,154 @@
+# JoyAI-Image
+
+JoyAI-Image is a unified multi-modal foundation model open-sourced by JD.com, supporting image understanding, text-to-image generation, and instruction-guided image editing.
+
+## Installation
+
+Before performing model inference and training, please install DiffSynth-Studio first.
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+For more information on installation, please refer to [Setup Dependencies](../Pipeline_Usage/Setup.md).
+
+## Quick Start
+
+Running the following code will load the [jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit) model for inference. VRAM management is enabled, the framework automatically controls parameter loading based on available VRAM, requiring a minimum of 4GB VRAM.
+
+```python
+from diffsynth.pipelines.joyai_image import JoyAIImagePipeline, ModelConfig
+import torch
+from PIL import Image
+from modelscope import dataset_snapshot_download
+
+# Download dataset
+dataset_snapshot_download(
+    dataset_id="DiffSynth-Studio/diffsynth_example_dataset",
+    local_dir="data/diffsynth_example_dataset",
+    allow_file_pattern="joyai_image/JoyAI-Image-Edit/*"
+)
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+
+pipe = JoyAIImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="transformer/transformer.pth", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/model*.safetensors", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="vae/Wan2.1_VAE.pth", **vram_config),
+    ],
+    processor_config=ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+# Use first sample from dataset
+dataset_base_path = "data/diffsynth_example_dataset/joyai_image/JoyAI-Image-Edit"
+prompt = "将裙子改为粉色"
+edit_image = Image.open(f"{dataset_base_path}/edit/image1.jpg").convert("RGB")
+
+output = pipe(
+    prompt=prompt,
+    edit_image=edit_image,
+    height=1024,
+    width=1024,
+    seed=0,
+    num_inference_steps=30,
+    cfg_scale=5.0,
+)
+
+output.save("output_joyai_edit_low_vram.png")
+```
+
+## Model Overview
+
+|Model ID|Inference|Low VRAM Inference|Full Training|Full Training Validation|LoRA Training|LoRA Training Validation|
+|-|-|-|-|-|-|-|
+|[jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_inference/JoyAI-Image-Edit.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_inference_low_vram/JoyAI-Image-Edit.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/full/JoyAI-Image-Edit.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/validate_full/JoyAI-Image-Edit.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/lora/JoyAI-Image-Edit.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/validate_lora/JoyAI-Image-Edit.py)|
+
+## Model Inference
+
+The model is loaded via `JoyAIImagePipeline.from_pretrained`, see [Loading Models](../Pipeline_Usage/Model_Inference.md#loading-models) for details.
+
+The input parameters for `JoyAIImagePipeline` inference include:
+
+* `prompt`: Text prompt describing the desired image editing effect.
+* `negative_prompt`: Negative prompt specifying what should not appear in the result, defaults to empty string.
+* `cfg_scale`: Classifier-free guidance scale factor, defaults to 5.0. Higher values make the output more closely follow the prompt.
+* `edit_image`: Image to be edited.
+* `denoising_strength`: Denoising strength controlling how much the input image is repainted, defaults to 1.0.
+* `height`: Height of the output image, defaults to 1024. Must be divisible by 16.
+* `width`: Width of the output image, defaults to 1024. Must be divisible by 16.
+* `seed`: Random seed for reproducibility. Set to `None` for random seed.
+* `max_sequence_length`: Maximum sequence length for the text encoder, defaults to 4096.
+* `num_inference_steps`: Number of inference steps, defaults to 30. More steps typically yield better quality.
+* `tiled`: Whether to enable tiling for reduced VRAM usage, defaults to False.
+* `tile_size`: Tile size, defaults to (30, 52).
+* `tile_stride`: Tile stride, defaults to (15, 26).
+* `shift`: Shift parameter for the scheduler, controlling the Flow Match scheduling curve, defaults to 4.0.
+* `progress_bar_cmd`: Progress bar display mode, defaults to tqdm.
+
+## Model Training
+
+Models in the joyai_image series are trained uniformly via `examples/joyai_image/model_training/train.py`. The script parameters include:
+
+* General Training Parameters
+    * Dataset Configuration
+        * `--dataset_base_path`: Root directory of the dataset.
+        * `--dataset_metadata_path`: Path to the dataset metadata file.
+        * `--dataset_repeat`: Number of dataset repeats per epoch.
+        * `--dataset_num_workers`: Number of processes per DataLoader.
+        * `--data_file_keys`: Field names to load from metadata, typically paths to image or video files, separated by `,`.
+    * Model Loading Configuration
+        * `--model_paths`: Paths to load models from, in JSON format.
+        * `--model_id_with_origin_paths`: Model IDs with original paths, separated by commas.
+        * `--extra_inputs`: Additional input parameters required by the model Pipeline, separated by `,`.
+        * `--fp8_models`: Models to load in FP8 format, currently only supported for models whose parameters are not updated by gradients.
+    * Basic Training Configuration
+        * `--learning_rate`: Learning rate.
+        * `--num_epochs`: Number of epochs.
+        * `--trainable_models`: Trainable models, e.g., `dit`, `vae`, `text_encoder`.
+        * `--find_unused_parameters`: Whether unused parameters exist in DDP training.
+        * `--weight_decay`: Weight decay magnitude.
+        * `--task`: Training task, defaults to `sft`.
+    * Output Configuration
+        * `--output_path`: Path to save the model.
+        * `--remove_prefix_in_ckpt`: Remove prefix in the model's state dict.
+        * `--save_steps`: Interval in training steps to save the model.
+    * LoRA Configuration
+        * `--lora_base_model`: Which model to add LoRA to.
+        * `--lora_target_modules`: Which layers to add LoRA to.
+        * `--lora_rank`: Rank of LoRA.
+        * `--lora_checkpoint`: Path to LoRA checkpoint.
+        * `--preset_lora_path`: Path to preset LoRA checkpoint for LoRA differential training.
+        * `--preset_lora_model`: Which model to integrate preset LoRA into, e.g., `dit`.
+    * Gradient Configuration
+        * `--use_gradient_checkpointing`: Whether to enable gradient checkpointing.
+        * `--use_gradient_checkpointing_offload`: Whether to offload gradient checkpointing to CPU memory.
+        * `--gradient_accumulation_steps`: Number of gradient accumulation steps.
+    * Resolution Configuration
+        * `--height`: Height of the image/video. Leave empty to enable dynamic resolution.
+        * `--width`: Width of the image/video. Leave empty to enable dynamic resolution.
+        * `--max_pixels`: Maximum pixel area, images larger than this will be scaled down during dynamic resolution.
+        * `--num_frames`: Number of frames for video (video generation models only).
+* JoyAI-Image Specific Parameters
+    * `--processor_path`: Path to the processor for processing text and image encoder inputs.
+    * `--initialize_model_on_cpu`: Whether to initialize models on CPU. By default, models are initialized on the accelerator device.
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
+```
+
+We provide recommended training scripts for each model, please refer to the table in "Model Overview" above. For guidance on writing model training scripts, see [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, see [Training Framework Overview](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/LTX-2.md

@@ -16,7 +16,7 @@ For more information about installation, please refer to [Installation Dependenc
 
 ## Quick Start
 
-Run the following code to quickly load the [Lightricks/LTX-2](https://www.modelscope.cn/models/Lightricks/LTX-2) model and perform inference. VRAM management has been enabled, and the framework will automatically control model parameter loading based on remaining VRAM. It can run with a minimum of 8GB VRAM.
+Run the following code to quickly load the [Lightricks/LTX-2.3](https://www.modelscope.cn/models/Lightricks/LTX-2.3) model and perform inference. VRAM management has been enabled, and the framework will automatically control model parameter loading based on remaining VRAM. It can run with a minimum of 8GB VRAM.
 
 ```python
 import torch
@@ -24,88 +24,36 @@ from diffsynth.pipelines.ltx2_audio_video import LTX2AudioVideoPipeline, ModelCo
 from diffsynth.utils.data.media_io_ltx2 import write_video_audio_ltx2
 
 vram_config = {
-    "offload_dtype": torch.float8_e5m2,
+    "offload_dtype": torch.bfloat16,
     "offload_device": "cpu",
-    "onload_dtype": torch.float8_e5m2,
-    "onload_device": "cpu",
-    "preparing_dtype": torch.float8_e5m2,
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cuda",
+    "preparing_dtype": torch.bfloat16,
     "preparing_device": "cuda",
     "computation_dtype": torch.bfloat16,
     "computation_device": "cuda",
 }
-"""
-Offical model repo: https://www.modelscope.cn/models/Lightricks/LTX-2
-Repackaged model repo: https://www.modelscope.cn/models/DiffSynth-Studio/LTX-2-Repackage
-For base models of LTX-2, offical checkpoint (with model config ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-dev.safetensors"))
-and repackaged checkpoints (with model config ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="*.safetensors")) are both supported.
-We have repackeged the official checkpoints in DiffSynth-Studio/LTX-2-Repackage repo to support separate loading of different submodules,
-and avoid redundant memory usage when users only want to use part of the model.
-"""
-# use the repackaged modelconfig from "DiffSynth-Studio/LTX-2-Repackage" to avoid redundant model loading
 pipe = LTX2AudioVideoPipeline.from_pretrained(
     torch_dtype=torch.bfloat16,
     device="cuda",
     model_configs=[
         ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized", origin_file_pattern="model-*.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="transformer.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="text_encoder_post_modules.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="video_vae_decoder.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="audio_vae_decoder.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="audio_vocoder.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="video_vae_encoder.safetensors", **vram_config),
-        ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-spatial-upscaler-x2-1.0.safetensors", **vram_config),
+        ModelConfig(model_id="Lightricks/LTX-2.3", origin_file_pattern="ltx-2.3-22b-dev.safetensors", **vram_config),
+        ModelConfig(model_id="Lightricks/LTX-2.3", origin_file_pattern="ltx-2.3-spatial-upscaler-x2-1.0.safetensors", **vram_config),
     ],
     tokenizer_config=ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized"),
-    stage2_lora_config=ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-distilled-lora-384.safetensors"),
-    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+    stage2_lora_config=ModelConfig(model_id="Lightricks/LTX-2.3", origin_file_pattern="ltx-2.3-22b-distilled-lora-384.safetensors"),
 )
-
-# use the following modelconfig if you want to initialize model from offical checkpoints from "Lightricks/LTX-2"
-# pipe = LTX2AudioVideoPipeline.from_pretrained(
-#     torch_dtype=torch.bfloat16,
-#     device="cuda",
-#     model_configs=[
-#         ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized", origin_file_pattern="model-*.safetensors", **vram_config),
-#         ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-dev.safetensors", **vram_config),
-#         ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-spatial-upscaler-x2-1.0.safetensors", **vram_config),
-#     ],
-#     tokenizer_config=ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized"),
-#     stage2_lora_config=ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-distilled-lora-384.safetensors"),
-#     vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
-# )
-
-prompt = "A girl is very happy, she is speaking: \"I enjoy working with Diffsynth-Studio, it's a perfect framework.\""
-negative_prompt = (
-    "blurry, out of focus, overexposed, underexposed, low contrast, washed out colors, excessive noise, "
-    "grainy texture, poor lighting, flickering, motion blur, distorted proportions, unnatural skin tones, "
-    "deformed facial features, asymmetrical face, missing facial features, extra limbs, disfigured hands, "
-    "wrong hand count, artifacts around text, inconsistent perspective, camera shake, incorrect depth of "
-    "field, background too sharp, background clutter, distracting reflections, harsh shadows, inconsistent "
-    "lighting direction, color banding, cartoonish rendering, 3D CGI look, unrealistic materials, uncanny "
-    "valley effect, incorrect ethnicity, wrong gender, exaggerated expressions, wrong gaze direction, "
-    "mismatched lip sync, silent or muted audio, distorted voice, robotic voice, echo, background noise, "
-    "off-sync audio, incorrect dialogue, added dialogue, repetitive speech, jittery movement, awkward "
-    "pauses, incorrect timing, unnatural transitions, inconsistent framing, tilted camera, flat lighting, "
-    "inconsistent tone, cinematic oversaturation, stylized filters, or AI artifacts."
-)
-height, width, num_frames = 512 * 2, 768 * 2, 121
+prompt = "Two cute orange cats, wearing boxing gloves, stand in a boxing ring and fight each other. They are punching each other fast and yelling: 'I will win!'"
+negative_prompt = pipe.default_negative_prompt["LTX-2.3"]
 video, audio = pipe(
     prompt=prompt,
     negative_prompt=negative_prompt,
     seed=43,
-    height=height,
-    width=width,
-    num_frames=num_frames,
-    tiled=True,
-    use_two_stage_pipeline=True,
-)
-write_video_audio_ltx2(
-    video=video,
-    audio=audio,
-    output_path='ltx2_twostage.mp4',
-    fps=24,
-    audio_sample_rate=24000,
+    height=1024, width=1536, num_frames=121,
+    tiled=True, use_two_stage_pipeline=True,
 )
+write_video_audio_ltx2(video=video, audio=audio, output_path='video.mp4', fps=24, audio_sample_rate=pipe.audio_vocoder.output_sampling_rate)
 ```
 
 ## Model Overview
@@ -217,7 +165,7 @@ LTX-2 series models are uniformly trained through [`examples/ltx2/model_training
 We have built a sample video dataset for your testing. You can download this dataset with the following command:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_video_dataset --local_dir ./data/example_video_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 We have written recommended training scripts for each model, please refer to the table in the "Model Overview" section above. For how to write model training scripts, please refer to [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, please refer to [Training Framework Detailed Explanation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/Qwen-Image.md

@@ -199,7 +199,7 @@ Qwen-Image series models are uniformly trained through [`examples/qwen_image/mod
 We have built a sample image dataset for your testing. You can download this dataset with the following command:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 We have written recommended training scripts for each model, please refer to the table in the "Model Overview" section above. For how to write model training scripts, please refer to [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, please refer to [Training Framework Detailed Explanation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/Stable-Diffusion-XL.md

@@ -0,0 +1,141 @@
+# Stable Diffusion XL
+
+Stable Diffusion XL (SDXL) is an open-source diffusion-based text-to-image generation model developed by Stability AI, supporting 1024x1024 resolution high-quality text-to-image generation with a dual text encoder (CLIP-L + CLIP-bigG) architecture.
+
+## Installation
+
+Before performing model inference and training, please install DiffSynth-Studio first.
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+For more information on installation, please refer to [Setup Dependencies](../Pipeline_Usage/Setup.md).
+
+## Quick Start
+
+Running the following code will quickly load the [stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0) model for inference. VRAM management is enabled, the framework automatically controls parameter loading based on available VRAM, requiring a minimum of 6GB VRAM.
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion_xl import StableDiffusionXLPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder_2/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer/"),
+    tokenizer_2_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer_2/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars",
+    negative_prompt="",
+    cfg_scale=5.0,
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+## Model Overview
+
+|Model ID|Inference|Low VRAM Inference|Full Training|Full Training Validation|LoRA Training|LoRA Training Validation|
+|-|-|-|-|-|-|-|
+|[stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_inference/stable-diffusion-xl-base-1.0.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_inference_low_vram/stable-diffusion-xl-base-1.0.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/full/stable-diffusion-xl-base-1.0.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/validate_full/stable-diffusion-xl-base-1.0.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/lora/stable-diffusion-xl-base-1.0.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/validate_lora/stable-diffusion-xl-base-1.0.py)|
+
+## Model Inference
+
+The model is loaded via `StableDiffusionXLPipeline.from_pretrained`, see [Loading Models](../Pipeline_Usage/Model_Inference.md#loading-models) for details.
+
+The input parameters for `StableDiffusionXLPipeline` inference include:
+
+* `prompt`: Text prompt.
+* `negative_prompt`: Negative prompt, defaults to an empty string.
+* `cfg_scale`: Classifier-Free Guidance scale factor, default 5.0.
+* `height`: Output image height, default 1024.
+* `width`: Output image width, default 1024.
+* `seed`: Random seed, defaults to a random value if not set.
+* `rand_device`: Noise generation device, defaults to "cpu".
+* `num_inference_steps`: Number of inference steps, default 50.
+* `guidance_rescale`: Guidance rescale factor, default 0.0.
+* `progress_bar_cmd`: Progress bar callback function.
+
+> `StableDiffusionXLPipeline` requires dual tokenizer configurations (`tokenizer_config` and `tokenizer_2_config`), corresponding to the CLIP-L and CLIP-bigG text encoders.
+
+## Model Training
+
+Models in the stable_diffusion_xl series are trained via `examples/stable_diffusion_xl/model_training/train.py`. The script parameters include:
+
+* General Training Parameters
+    * Dataset Configuration
+        * `--dataset_base_path`: Root directory of the dataset.
+        * `--dataset_metadata_path`: Path to the dataset metadata file.
+        * `--dataset_repeat`: Number of dataset repeats per epoch.
+        * `--dataset_num_workers`: Number of processes per DataLoader.
+        * `--data_file_keys`: Field names to load from metadata, typically paths to image or video files, separated by `,`.
+    * Model Loading Configuration
+        * `--model_paths`: Paths to load models from, in JSON format.
+        * `--model_id_with_origin_paths`: Model IDs with original paths, separated by commas.
+        * `--extra_inputs`: Additional input parameters required by the model Pipeline, separated by `,`.
+        * `--fp8_models`: Models to load in FP8 format, currently only supported for models whose parameters are not updated by gradients.
+    * Basic Training Configuration
+        * `--learning_rate`: Learning rate.
+        * `--num_epochs`: Number of epochs.
+        * `--trainable_models`: Trainable models, e.g., `dit`, `vae`, `text_encoder`.
+        * `--find_unused_parameters`: Whether unused parameters exist in DDP training.
+        * `--weight_decay`: Weight decay magnitude.
+        * `--task`: Training task, defaults to `sft`.
+    * Output Configuration
+        * `--output_path`: Path to save the model.
+        * `--remove_prefix_in_ckpt`: Remove prefix in the model's state dict.
+        * `--save_steps`: Interval in training steps to save the model.
+    * LoRA Configuration
+        * `--lora_base_model`: Which model to add LoRA to.
+        * `--lora_target_modules`: Which layers to add LoRA to.
+        * `--lora_rank`: Rank of LoRA.
+        * `--lora_checkpoint`: Path to LoRA checkpoint.
+        * `--preset_lora_path`: Path to preset LoRA checkpoint for LoRA differential training.
+        * `--preset_lora_model`: Which model to integrate preset LoRA into, e.g., `dit`.
+    * Gradient Configuration
+        * `--use_gradient_checkpointing`: Whether to enable gradient checkpointing.
+        * `--use_gradient_checkpointing_offload`: Whether to offload gradient checkpointing to CPU memory.
+        * `--gradient_accumulation_steps`: Number of gradient accumulation steps.
+    * Resolution Configuration
+        * `--height`: Height of the image/video. Leave empty to enable dynamic resolution.
+        * `--width`: Width of the image/video. Leave empty to enable dynamic resolution.
+        * `--max_pixels`: Maximum pixel area, images larger than this will be scaled down during dynamic resolution.
+        * `--num_frames`: Number of frames for video (video generation models only).
+* Stable Diffusion XL Specific Parameters
+    * `--tokenizer_path`: Path to the first tokenizer.
+    * `--tokenizer_2_path`: Path to the second tokenizer, defaults to `stabilityai/stable-diffusion-xl-base-1.0:tokenizer_2/`.
+
+Example dataset download:
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "stable_diffusion_xl/*" --local_dir ./data/diffsynth_example_dataset
+```
+
+[stable-diffusion-xl-base-1.0 training scripts](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/lora/stable-diffusion-xl-base-1.0.sh)
+
+We provide recommended training scripts for each model, please refer to the table in "Model Overview" above. For guidance on writing model training scripts, see [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, see [Training Framework Overview](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/Stable-Diffusion.md

@@ -0,0 +1,138 @@
+# Stable Diffusion
+
+Stable Diffusion is an open-source diffusion-based text-to-image generation model developed by Stability AI, supporting 512x512 resolution text-to-image generation.
+
+## Installation
+
+Before performing model inference and training, please install DiffSynth-Studio first.
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+For more information on installation, please refer to [Setup Dependencies](../Pipeline_Usage/Setup.md).
+
+## Quick Start
+
+Running the following code will quickly load the [AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5) model for inference. VRAM management is enabled, the framework automatically controls parameter loading based on available VRAM, requiring a minimum of 2GB VRAM.
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion import StableDiffusionPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars, high quality, detailed",
+    negative_prompt="blurry, low quality, deformed",
+    cfg_scale=7.5,
+    height=512,
+    width=512,
+    seed=42,
+    rand_device="cuda",
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+## Model Overview
+
+|Model ID|Inference|Low VRAM Inference|Full Training|Full Training Validation|LoRA Training|LoRA Training Validation|
+|-|-|-|-|-|-|-|
+|[AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_inference/stable-diffusion-v1-5.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_inference_low_vram/stable-diffusion-v1-5.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/full/stable-diffusion-v1-5.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/validate_full/stable-diffusion-v1-5.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/lora/stable-diffusion-v1-5.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/validate_lora/stable-diffusion-v1-5.py)|
+
+## Model Inference
+
+The model is loaded via `StableDiffusionPipeline.from_pretrained`, see [Loading Models](../Pipeline_Usage/Model_Inference.md#loading-models) for details.
+
+The input parameters for `StableDiffusionPipeline` inference include:
+
+* `prompt`: Text prompt.
+* `negative_prompt`: Negative prompt, defaults to an empty string.
+* `cfg_scale`: Classifier-Free Guidance scale factor, default 7.5.
+* `height`: Output image height, default 512.
+* `width`: Output image width, default 512.
+* `seed`: Random seed, defaults to a random value if not set.
+* `rand_device`: Noise generation device, defaults to "cpu".
+* `num_inference_steps`: Number of inference steps, default 50.
+* `eta`: DDIM scheduler eta parameter, default 0.0.
+* `guidance_rescale`: Guidance rescale factor, default 0.0.
+* `progress_bar_cmd`: Progress bar callback function.
+
+## Model Training
+
+Models in the stable_diffusion series are trained via `examples/stable_diffusion/model_training/train.py`. The script parameters include:
+
+* General Training Parameters
+    * Dataset Configuration
+        * `--dataset_base_path`: Root directory of the dataset.
+        * `--dataset_metadata_path`: Path to the dataset metadata file.
+        * `--dataset_repeat`: Number of dataset repeats per epoch.
+        * `--dataset_num_workers`: Number of processes per DataLoader.
+        * `--data_file_keys`: Field names to load from metadata, typically paths to image or video files, separated by `,`.
+    * Model Loading Configuration
+        * `--model_paths`: Paths to load models from, in JSON format.
+        * `--model_id_with_origin_paths`: Model IDs with original paths, separated by commas.
+        * `--extra_inputs`: Additional input parameters required by the model Pipeline, separated by `,`.
+        * `--fp8_models`: Models to load in FP8 format, currently only supported for models whose parameters are not updated by gradients.
+    * Basic Training Configuration
+        * `--learning_rate`: Learning rate.
+        * `--num_epochs`: Number of epochs.
+        * `--trainable_models`: Trainable models, e.g., `dit`, `vae`, `text_encoder`.
+        * `--find_unused_parameters`: Whether unused parameters exist in DDP training.
+        * `--weight_decay`: Weight decay magnitude.
+        * `--task`: Training task, defaults to `sft`.
+    * Output Configuration
+        * `--output_path`: Path to save the model.
+        * `--remove_prefix_in_ckpt`: Remove prefix in the model's state dict.
+        * `--save_steps`: Interval in training steps to save the model.
+    * LoRA Configuration
+        * `--lora_base_model`: Which model to add LoRA to.
+        * `--lora_target_modules`: Which layers to add LoRA to.
+        * `--lora_rank`: Rank of LoRA.
+        * `--lora_checkpoint`: Path to LoRA checkpoint.
+        * `--preset_lora_path`: Path to preset LoRA checkpoint for LoRA differential training.
+        * `--preset_lora_model`: Which model to integrate preset LoRA into, e.g., `dit`.
+    * Gradient Configuration
+        * `--use_gradient_checkpointing`: Whether to enable gradient checkpointing.
+        * `--use_gradient_checkpointing_offload`: Whether to offload gradient checkpointing to CPU memory.
+        * `--gradient_accumulation_steps`: Number of gradient accumulation steps.
+    * Resolution Configuration
+        * `--height`: Height of the image/video. Leave empty to enable dynamic resolution.
+        * `--width`: Width of the image/video. Leave empty to enable dynamic resolution.
+        * `--max_pixels`: Maximum pixel area, images larger than this will be scaled down during dynamic resolution.
+        * `--num_frames`: Number of frames for video (video generation models only).
+* Stable Diffusion Specific Parameters
+    * `--tokenizer_path`: Tokenizer path, defaults to `AI-ModelScope/stable-diffusion-v1-5:tokenizer/`.
+
+Example dataset download:
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "stable_diffusion/*" --local_dir ./data/diffsynth_example_dataset
+```
+
+[stable-diffusion-v1-5 training scripts](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/lora/stable-diffusion-v1-5.sh)
+
+We provide recommended training scripts for each model, please refer to the table in "Model Overview" above. For guidance on writing model training scripts, see [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, see [Training Framework Overview](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/Wan.md

@@ -104,41 +104,43 @@ graph LR;
 
 </details>
 
-| Model ID | Extra Parameters | Inference | Full Training | Validation After Full Training | LoRA Training | Validation After LoRA Training |
-| - | - | - | - | - | - | - |
-| [Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B) | | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py) |
-| [Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B) | | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py) |
-| [Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py) |
-| [Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py) |
-| [Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P) | `input_image`, `end_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py) |
-| [iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview) | `vace_control_video`, `vace_reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py) |
-| [Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B) | `vace_control_video`, `vace_reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py) |
-| [Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B) | `vace_control_video`, `vace_reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py) |
-| [PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP) | `input_image`, `end_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py) |
-| [PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control) | `control_video` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py) |
-| [PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP) | `input_image`, `end_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py) |
-| [PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control) | `control_video` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py) |
-| [PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control) | `control_video`, `reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py) |
-| [PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control) | `control_video`, `reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py) |
-| [PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP) | `input_image`, `end_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py) |
-| [PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP) | `input_image`, `end_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py) |
-| [PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera) | `control_camera_video`, `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py) |
-| [PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera) | `control_camera_video`, `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py) |
-| [DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1) | `motion_bucket_id` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py) |
-| [krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video) | | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/krea-realtime-video.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/krea-realtime-video.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/krea-realtime-video.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/krea-realtime-video.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py) |
-| [meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video) | `longcat_video` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/LongCat-Video.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/LongCat-Video.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/LongCat-Video.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/LongCat-Video.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/LongCat-Video.py) |
-| [ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B) | `vap_video`, `vap_prompt` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py) |
-| [Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B) | | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py) |
-| [Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py) |
-| [Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py) |
-| [Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B) | `input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py) |
-| [Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B) | `input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py) |
-| [PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B) | `vace_control_video`, `vace_reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py) |
-| [PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP) | `input_image`, `end_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py) |
-| [PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control) | `control_video`, `reference_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py) |
-| [PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera) | `control_camera_video`, `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py) |
-| [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-360p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-360P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-360P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py) |
-| [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-720p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-720P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-720P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py) |
+| Model ID | Extra Inputs | Inference | Low VRAM Inference | Full Training | Validation After Full Training | LoRA Training | Validation After LoRA Training |
+|-|-|-|-|-|-|-|-|
+|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
+|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
+|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
+|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
+|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
+|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
+|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
+|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
+|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
+|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
+|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
+|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
+|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
+|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
+|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
+|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
+|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
+|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
+|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
+|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
+|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
+|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
+|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
+|[openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py)|
+|[openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py)|
+|[Wan-AI/WanToDance-14B (global model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-global.py)|
+|[Wan-AI/WanToDance-14B (local model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-local.py)|
 
 * FP8 Precision Training: [doc](../Training/FP8_Precision.md), [code](https://github.com/modelscope/DiffSynth-Studio/tree/main/examples/wanvideo/model_training/special/fp8_training/)
 * Two-stage Split Training: [doc](../Training/Split_Training.md), [code](https://github.com/modelscope/DiffSynth-Studio/tree/main/examples/wanvideo/model_training/special/split_training/)
@@ -203,6 +205,50 @@ Input parameters for `WanVideoPipeline` inference include:
 
 If VRAM is insufficient, please enable [VRAM Management](../Pipeline_Usage/VRAM_management.md). We provide recommended low VRAM configurations for each model in the example code, see the table in the "Model Overview" section above.
 
+### Multi-GPU Parallel Acceleration
+
+To enable multi-GPU parallel acceleration, please install `flash_attn` and `xfuser`:
+
+```shell
+pip install flash-attn --no-build-isolation
+pip install xfuser
+```
+
+Please modify your code as follows ([example code](https://github.com/modelscope/DiffSynth-Studio/tree/main/examples/wanvideo/acceleration/unified_sequence_parallel.py)):
+
+```diff
+import torch
+from PIL import Image
+from diffsynth.utils.data import save_video, VideoData
+from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
++ import torch.distributed as dist
+
+pipe = WanVideoPipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
++   use_usp=True,
+    model_configs=[
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="models_t5_umt5-xxl-enc-bf16.pth"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="Wan2.1_VAE.pth"),
+    ],
+    tokenizer_config=ModelConfig(model_id="Wan-AI/Wan2.1-T2V-1.3B", origin_file_pattern="google/umt5-xxl/"),
+)
+video = pipe(
+    prompt="An astronaut in a spacesuit rides a mechanical horse across the Martian surface, facing the camera. The red, desolate terrain stretches into the distance, dotted with massive craters and unusual rock formations. The mechanical horse moves with steady strides, kicking up faint dust, embodying a perfect fusion of futuristic technology and primal exploration. The astronaut holds a control device, with a determined gaze, as if pioneering new frontiers for humanity. Against a backdrop of the deep cosmos and the blue Earth, the scene is both sci-fi and hopeful, evoking imagination about future interstellar life.",
+    negative_prompt="oversaturated colors, overexposed, static, blurry details, subtitles, style, artwork, painting, still image, overall gray tone, worst quality, low quality, JPEG compression artifacts, ugly, malformed, extra fingers, poorly drawn hands, poorly drawn face, deformed, disfigured, malformed limbs, fused fingers, frozen frame, cluttered background, three legs, crowd in background, walking backwards",
+    seed=0, tiled=True,
+)
++ if dist.get_rank() == 0:
++   save_video(video, "video1.mp4", fps=15, quality=5)
+```
+
+When running multi-GPU parallel inference, please use `torchrun`, where `--nproc_per_node` specifies the number of GPUs:
+
+```shell
+torchrun --nproc_per_node=8 examples/wanvideo/acceleration/unified_sequence_parallel.py
+```
+
 ## Model Training
 
 Wan series models are uniformly trained through [`examples/wanvideo/model_training/train.py`](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/train.py), and the script parameters include:
@@ -253,7 +299,7 @@ Wan series models are uniformly trained through [`examples/wanvideo/model_traini
 We have built a sample video dataset for your testing. You can download this dataset with the following command:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_video_dataset --local_dir ./data/example_video_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 We have written recommended training scripts for each model, please refer to the table in the "Model Overview" section above. For how to write model training scripts, please refer to [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, please refer to [Training Framework Detailed Explanation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Model_Details/Z-Image.md

@@ -134,7 +134,7 @@ Z-Image series models are uniformly trained through [`examples/z_image/model_tra
 We have built a sample image dataset for your testing. You can download this dataset with the following command:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 We have written recommended training scripts for each model, please refer to the table in the "Model Overview" section above. For how to write model training scripts, please refer to [Model Training](../Pipeline_Usage/Model_Training.md); for more advanced training algorithms, please refer to [Training Framework Detailed Explanation](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/en/Training/).

docs/en/Pipeline_Usage/Accelerated_Inference.md

@@ -0,0 +1,94 @@
+# Inference Acceleration
+
+The denoising process of diffusion models is typically time-consuming. To improve inference speed, various acceleration techniques can be applied, including lossless acceleration solutions such as multi-GPU parallel inference and computation graph compilation, as well as lossy acceleration solutions like Cache and quantization.
+
+Currently, most diffusion models are built on Diffusion Transformers (DiT), and efficient attention mechanisms are also a common acceleration method. DiffSynth-Studio currently supports certain lossless acceleration inference features. This section focuses on introducing acceleration methods from two dimensions: multi-GPU parallel inference and computation graph compilation.
+
+## Efficient Attention Mechanisms
+
+For details on the acceleration of attention mechanisms, please refer to [Attention Mechanism Implementation](../API_Reference/core/attention.md).
+
+## Multi-GPU Parallel Inference
+
+DiffSynth-Studio adopts a multi-GPU inference solution using Unified Sequence Parallel (USP). It splits the token sequence in the DiT across multiple GPUs for parallel processing. The underlying implementation is based on [xDiT](https://github.com/xdit-project/xDiT). Please note that unified sequence parallelism introduces additional communication overhead, so the actual speedup ratio is usually lower than the number of GPUs.
+
+Currently, DiffSynth-Studio supports unified sequence parallel acceleration for the [Wan](../Model_Details/Wan.md) and [MOVA](../Model_Details/Wan.md) models.
+
+First, install the `xDiT` dependency.
+
+```bash
+pip install "xfuser[flash-attn]>=0.4.3"
+```
+
+Then, use `torchrun` to launch multi-GPU inference.
+
+```bash
+torchrun --standalone --nproc_per_node=8 examples/wanvideo/acceleration/unified_sequence_parallel.py
+```
+
+When building the pipeline, simply configure `use_usp=True` to enable USP parallel inference. A code example is shown below.
+
+```python
+import torch
+from PIL import Image
+from diffsynth.utils.data import save_video
+from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
+import torch.distributed as dist
+
+pipe = WanVideoPipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    use_usp=True,
+    model_configs=[
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="models_t5_umt5-xxl-enc-bf16.pth"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="Wan2.1_VAE.pth"),
+    ],
+    tokenizer_config=ModelConfig(model_id="Wan-AI/Wan2.1-T2V-1.3B", origin_file_pattern="google/umt5-xxl/"),
+)
+
+# Text-to-video
+video = pipe(
+    prompt="一名宇航员身穿太空服，面朝镜头骑着一匹机械马在火星表面驰骋。红色的荒凉地表延伸至远方，点缀着巨大的陨石坑和奇特的岩石结构。机械马的步伐稳健，扬起微弱的尘埃，展现出未来科技与原始探索的完美结合。宇航员手持操控装置，目光坚定，仿佛正在开辟人类的新疆域。背景是深邃的宇宙和蔚蓝的地球，画面既科幻又充满希望，让人不禁畅想未来的星际生活。",
+    negative_prompt="色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走",
+    seed=0, tiled=True,
+)
+if dist.get_rank() == 0:
+    save_video(video, "video1.mp4", fps=15, quality=5)
+```
+
+## Computation Graph Compilation
+
+PyTorch 2.0 provides an automatic computation graph compilation interface, [torch.compile](https://docs.pytorch.org/tutorials/intermediate/torch_compile_tutorial.html), which can just-in-time (JIT) compile PyTorch code into optimized kernels, thereby improving execution speed. Since the inference time of diffusion models is concentrated in the multi-step denoising phase of the DiT, and the DiT is primarily stacked with basic blocks, DiffSynth's compile feature uses a [regional compilation](https://docs.pytorch.org/tutorials/recipes/regional_compilation.html) strategy targeting only the basic Transformer blocks to reduce compilation time.
+
+### Compile Usage Example
+
+Compared to standard inference, you simply need to execute `pipe.compile_pipeline()` before calling the pipeline to enable compilation acceleration. For the specific function definition, please refer to the [source code](https://github.com/modelscope/DiffSynth-Studio/blob/166e6d2d38764209f66a74dd0fe468226536ad0f/diffsynth/diffusion/base_pipeline.py#L342).
+
+The input parameters for `compile_pipeline` consist mainly of two types.
+
+The first type is the compiled model parameters, `compile_models`. Taking the Qwen-Image Pipeline as an example, if you only want to compile the DiT model, you can keep this parameter empty. If you need to additionally compile models like the VAE, you can pass `compile_models=["vae", "dit"]`. Aside from DiT, all other models use a full-graph compilation strategy, meaning the model's forward function is completely compiled into a computation graph.
+
+The second type is the compilation strategy parameters. This covers `mode`, `dynamic`, `fullgraph`, and other custom options. These parameters are directly passed to the `torch.compile` interface. If you are not deeply familiar with the specific mechanics of these parameters, it is recommended to keep the default settings.
+
+  * `mode` specifies the compilation mode, including `"default"`, `"reduce-overhead"`, `"max-autotune"`, and `"max-autotune-no-cudagraphs"`. Because cudagraph has stricter requirements on computation graphs (for example, it might need to be used in conjunction with `torch.compiler.cudagraph_mark_step_begin()`), the `"reduce-overhead"` and `"max-autotune"` modes might fail to compile.
+  * `dynamic` determines whether to enable dynamic shapes. For most generative models, modifying the prompt, enabling CFG, or adjusting the resolution will change the shape of the input tensors to the computation graph. Setting `dynamic=True` will increase the compilation time of the first run, but it supports dynamic shapes, meaning no recompilation is needed when shapes change. When set to `dynamic=False`, the first compilation is faster, but any operation that alters the input shape will trigger a recompilation. For most scenarios, setting it to `dynamic=True` is recommended.
+  * `fullgraph`, when set to `True`, makes the underlying system attempt to compile the target model into a single computation graph, throwing an error if it fails. When set to `False`, the underlying system will set breakpoints where connections cannot be made, compiling the model into multiple independent computation graphs. Developers can set it to `True` to optimize compilation performance, but regular users are advised to only use `False`.
+  * For other parameter configurations, please consult the [API documentation](https://docs.pytorch.org/docs/stable/generated/torch.compile.html).
+
+### Compile Feature Developer Documentation
+
+If you need to provide compile support for a newly integrated pipeline, you should configure the `compilable_models` attribute in the pipeline to specify the default models to compile. For the DiT model class of that pipeline, you also need to configure `_repeated_blocks` to specify the types of basic blocks that will participate in regional compilation.
+
+Taking Qwen-Image as an example, its pipeline configuration is as follows:
+
+```python
+self.compilable_models = ["dit"]
+```
+
+Its DiT configuration is as follows:
+
+```python
+class QwenImageDiT(torch.nn.Module):
+    _repeated_blocks = ["QwenImageTransformerBlock"]
+```

docs/en/Pipeline_Usage/Model_Training.md

@@ -69,25 +69,11 @@ We have built sample datasets for your testing. To understand how the universal
 
 <details>
 
-<summary>Sample Image Dataset</summary>
+<summary>Sample Dataset</summary>
 
 > ```shell
-> modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+> modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 > ```
-> 
-> Applicable to training of image generation models such as Qwen-Image and FLUX.
-
-</details>
-
-<details>
-
-<summary>Sample Video Dataset</summary>
-
-> ```shell
-> modelscope download --dataset DiffSynth-Studio/example_video_dataset --local_dir ./data/example_video_dataset
-> ```
-> 
-> Applicable to training of video generation models such as Wan.
 
 </details>
 
@@ -123,7 +109,6 @@ Similar to [model loading during inference](../Pipeline_Usage/Model_Inference.md
 
 <details>
 
-<details>
 
 <summary>Load models from local file paths</summary>
 
@@ -244,4 +229,119 @@ accelerate launch --config_file examples/qwen_image/model_training/full/accelera
 * The training framework does not support batch size > 1. The reasons are complex. See [Q&A: Why doesn't the training framework support batch size > 1?](../QA.md#why-doesnt-the-training-framework-support-batch-size--1)
 * Some models contain redundant parameters. For example, the text encoding part of the last layer of Qwen-Image's DiT part. When training these models, `--find_unused_parameters` needs to be set to avoid errors in multi-GPU training. For compatibility with community models, we do not intend to remove these redundant parameters.
 * The loss function value of Diffusion models has little relationship with actual effects. Therefore, we do not record loss function values during training. We recommend setting `--num_epochs` to a sufficiently large value, testing while training, and manually closing the training program after the effect converges.
-* `--use_gradient_checkpointing` is usually enabled unless GPU VRAM is sufficient; `--use_gradient_checkpointing_offload` is enabled as needed. See [`diffsynth.core.gradient`](../API_Reference/core/gradient.md) for details.
+* `--use_gradient_checkpointing` is usually enabled unless GPU VRAM is sufficient; `--use_gradient_checkpointing_offload` is enabled as needed. See [`diffsynth.core.gradient`](../API_Reference/core/gradient.md) for details.
+
+## Low VRAM Training
+
+If you want to complete LoRA model training on GPU with low vram, you can combine [Two-Stage Split Training](../Training/Split_Training.md) with `deepspeed_zero3_offload` training. First, split the preprocessing steps into the first stage and store the computed results onto the hard disk. Second, read these results from the disk and train the denoising model. By using `deepspeed_zero3_offload`, the training parameters and optimizer states are offloaded to the CPU or disk. We provide examples for some models, primarily by specifying the `deepspeed` configuration via `--config_file`.
+
+Please note that the `deepspeed_zero3_offload` mode is incompatible with PyTorch's native gradient checkpointing mechanism. To address this, we have adapted the `checkpointing` interface of `deepspeed`. Users need to fill the `activation_checkpointing` field in the `deepspeed` configuration to enable gradient checkpointing.
+
+Below is the script for low VRAM model training for the Qwen-Image model:
+
+```shell
+accelerate launch examples/qwen_image/model_training/train.py \
+  --dataset_base_path data/example_image_dataset \
+  --dataset_metadata_path data/example_image_dataset/metadata.csv \
+  --max_pixels 1048576 \
+  --dataset_repeat 1 \
+  --model_id_with_origin_paths "Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \
+  --learning_rate 1e-4 \
+  --num_epochs 5 \
+  --remove_prefix_in_ckpt "pipe.dit." \
+  --output_path "./models/train/Qwen-Image_lora-splited-cache" \
+  --lora_base_model "dit" \
+  --lora_target_modules "to_q,to_k,to_v,add_q_proj,add_k_proj,add_v_proj,to_out.0,to_add_out,img_mlp.net.2,img_mod.1,txt_mlp.net.2,txt_mod.1" \
+  --lora_rank 32 \
+  --task "sft:data_process" \
+  --use_gradient_checkpointing \
+  --dataset_num_workers 8 \
+  --find_unused_parameters
+
+accelerate launch --config_file examples/qwen_image/model_training/special/low_vram_training/deepspeed_zero3_cpuoffload.yaml examples/qwen_image/model_training/train.py \
+  --dataset_base_path "./models/train/Qwen-Image_lora-splited-cache" \
+  --max_pixels 1048576 \
+  --dataset_repeat 50 \
+  --model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors" \
+  --learning_rate 1e-4 \
+  --num_epochs 5 \
+  --remove_prefix_in_ckpt "pipe.dit." \
+  --output_path "./models/train/Qwen-Image_lora" \
+  --lora_base_model "dit" \
+  --lora_target_modules "to_q,to_k,to_v,add_q_proj,add_k_proj,add_v_proj,to_out.0,to_add_out,img_mlp.net.2,img_mod.1,txt_mlp.net.2,txt_mod.1" \
+  --lora_rank 32 \
+  --task "sft:train" \
+  --use_gradient_checkpointing \
+  --dataset_num_workers 8 \
+  --find_unused_parameters \
+  --initialize_model_on_cpu
+```
+
+The configurations for `accelerate` and `deepspeed` are as follows:
+
+```yaml
+compute_environment: LOCAL_MACHINE
+debug: true
+deepspeed_config:
+  deepspeed_config_file: examples/qwen_image/model_training/special/low_vram_training/ds_z3_cpuoffload.json
+  zero3_init_flag: true
+distributed_type: DEEPSPEED
+downcast_bf16: 'no'
+enable_cpu_affinity: false
+machine_rank: 0
+main_training_function: main
+num_machines: 1
+num_processes: 1
+rdzv_backend: static
+same_network: true
+tpu_env: []
+tpu_use_cluster: false
+tpu_use_sudo: false
+use_cpu: false
+```
+
+```json
+{
+    "fp16": {
+        "enabled": "auto",
+        "loss_scale": 0,
+        "loss_scale_window": 1000,
+        "initial_scale_power": 16,
+        "hysteresis": 2,
+        "min_loss_scale": 1
+    },
+    "bf16": {
+        "enabled": "auto"
+    },
+    "zero_optimization": {
+        "stage": 3,
+        "offload_optimizer": {
+            "device": "cpu",
+            "pin_memory": true
+        },
+        "offload_param": {
+            "device": "cpu",
+            "pin_memory": true
+        },
+        "overlap_comm": false,
+        "contiguous_gradients": true,
+        "sub_group_size": 1e9,
+        "reduce_bucket_size": 5e7,
+        "stage3_prefetch_bucket_size": 5e7,
+        "stage3_param_persistence_threshold": 1e5,
+        "stage3_max_live_parameters": 1e8,
+        "stage3_max_reuse_distance": 1e8,
+        "stage3_gather_16bit_weights_on_model_save": true
+    },
+    "activation_checkpointing": {
+        "partition_activations": false,
+        "cpu_checkpointing": false,
+        "contiguous_memory_optimization": false
+    },
+    "gradient_accumulation_steps": "auto",
+    "gradient_clipping": "auto",
+    "train_batch_size": "auto",
+    "train_micro_batch_size_per_gpu": "auto",
+    "wall_clock_breakdown": false
+}
+```

docs/en/Training/Direct_Distill.md

@@ -77,7 +77,7 @@ distill_qwen/image.jpg,"精致肖像，水下少女，蓝裙飘逸，发丝轻
 This sample dataset can be downloaded directly:
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 Then start LoRA distillation accelerated training:

docs/en/index.rst

@@ -13,6 +13,7 @@ Welcome to DiffSynth-Studio's Documentation
 
    Pipeline_Usage/Setup
    Pipeline_Usage/Model_Inference
+   Pipeline_Usage/Accelerated_Inference
    Pipeline_Usage/VRAM_management
    Pipeline_Usage/Model_Training
    Pipeline_Usage/Environment_Variables
@@ -29,6 +30,10 @@ Welcome to DiffSynth-Studio's Documentation
    Model_Details/Z-Image
    Model_Details/Anima
    Model_Details/LTX-2
+   Model_Details/ERNIE-Image
+   Model_Details/JoyAI-Image
+   Model_Details/Stable-Diffusion
+   Model_Details/Stable-Diffusion-XL
 
 .. toctree::
    :maxdepth: 2

docs/zh/Model_Details/Anima.md

@@ -133,7 +133,7 @@ Anima 系列模型统一通过 [`examples/anima/model_training/train.py`](https:
 我们构建了一个样例图像数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/ERNIE-Image.md

@@ -0,0 +1,134 @@
+# ERNIE-Image
+
+ERNIE-Image 是百度推出的拥有 8B 参数的图像生成模型，具有紧凑高效的架构和出色的指令跟随能力。基于 8B DiT 主干网络，其在某些场景下的性能可与 20B 以上的更大模型相媲美，同时保持了良好的参数效率。该模型在指令理解与执行、文本生成（如英文/中文/日文）以及整体稳定性方面提供了较为可靠的表现。
+
+## 安装
+
+在使用本项目进行模型推理和训练前，请先安装 DiffSynth-Studio。
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+更多关于安装的信息，请参考[安装依赖](../Pipeline_Usage/Setup.md)。
+
+## 快速开始
+
+运行以下代码可以快速加载 [PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 3G 显存即可运行。
+
+```python
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("output.jpg")
+```
+
+## 模型总览
+
+|模型 ID|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|
+|[PaddlePaddle/ERNIE-Image](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference/ERNIE-Image.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference_low_vram/ERNIE-Image.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/full/ERNIE-Image.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/validate_full/ERNIE-Image.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/lora/ERNIE-Image.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/validate_lora/ERNIE-Image.py)|
+|[PaddlePaddle/ERNIE-Image-Turbo](https://www.modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference/ERNIE-Image-Turbo.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_inference_low_vram/ERNIE-Image-Turbo.py)|—|—|—|—|
+
+## 模型推理
+
+模型通过 `ErnieImagePipeline.from_pretrained` 加载，详见[加载模型](../Pipeline_Usage/Model_Inference.md#加载模型)。
+
+`ErnieImagePipeline` 推理的输入参数包括：
+
+* `prompt`: 提示词，描述画面中出现的内容。
+* `negative_prompt`: 负向提示词，描述画面中不应该出现的内容，默认值为 `""`。
+* `cfg_scale`: Classifier-free guidance 的参数，默认值为 4.0。
+* `height`: 图像高度，需保证高度为 16 的倍数，默认值为 1024。
+* `width`: 图像宽度，需保证宽度为 16 的倍数，默认值为 1024。
+* `seed`: 随机种子。默认为 `None`，即完全随机。
+* `rand_device`: 生成随机高斯噪声矩阵的计算设备，默认为 `"cuda"`。当设置为 `cuda` 时，在不同 GPU 上会导致不同的生成结果。
+* `num_inference_steps`: 推理步数，默认值为 50。
+
+如果显存不足，请开启[显存管理](../Pipeline_Usage/VRAM_management.md)，我们在示例代码中提供了每个模型推荐的低显存配置，详见前文"模型总览"中的表格。
+
+## 模型训练
+
+ERNIE-Image 系列模型统一通过 [`examples/ernie_image/model_training/train.py`](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/ernie_image/model_training/train.py) 进行训练，脚本的参数包括：
+
+* 通用训练参数
+    * 数据集基础配置
+        * `--dataset_base_path`: 数据集的根目录。
+        * `--dataset_metadata_path`: 数据集的元数据文件路径。
+        * `--dataset_repeat`: 每个 epoch 中数据集重复的次数。
+        * `--dataset_num_workers`: 每个 Dataloader 的进程数量。
+        * `--data_file_keys`: 元数据中需要加载的字段名称，通常是图像或视频文件的路径，以 `,` 分隔。
+    * 模型加载配置
+        * `--model_paths`: 要加载的模型路径。JSON 格式。
+        * `--model_id_with_origin_paths`: 带原始路径的模型 ID，例如 `"PaddlePaddle/ERNIE-Image:transformer/diffusion_pytorch_model*.safetensors"`。用逗号分隔。
+        * `--extra_inputs`: 模型 Pipeline 所需的额外输入参数，以 `,` 分隔。
+        * `--fp8_models`：以 FP8 格式加载的模型，目前仅支持参数不被梯度更新的模型。
+    * 训练基础配置
+        * `--learning_rate`: 学习率。
+        * `--num_epochs`: 轮数（Epoch）。
+        * `--trainable_models`: 可训练的模型，例如 `dit`、`vae`、`text_encoder`。
+        * `--find_unused_parameters`: DDP 训练中是否存在未使用的参数。
+        * `--weight_decay`：权重衰减大小。
+        * `--task`: 训练任务，默认为 `sft`。
+    * 输出配置
+        * `--output_path`: 模型保存路径。
+        * `--remove_prefix_in_ckpt`: 在模型文件的 state dict 中移除前缀。
+        * `--save_steps`: 保存模型的训练步数间隔。
+    * LoRA 配置
+        * `--lora_base_model`: LoRA 添加到哪个模型上。
+        * `--lora_target_modules`: LoRA 添加到哪些层上。
+        * `--lora_rank`: LoRA 的秩（Rank）。
+        * `--lora_checkpoint`: LoRA 检查点的路径。
+        * `--preset_lora_path`: 预置 LoRA 检查点路径，用于 LoRA 差分训练。
+        * `--preset_lora_model`: 预置 LoRA 融入的模型，例如 `dit`。
+    * 梯度配置
+        * `--use_gradient_checkpointing`: 是否启用 gradient checkpointing。
+        * `--use_gradient_checkpointing_offload`: 是否将 gradient checkpointing 卸载到内存中。
+        * `--gradient_accumulation_steps`: 梯度累积步数。
+    * 分辨率配置
+        * `--height`: 图像的高度。留空启用动态分辨率。
+        * `--width`: 图像的宽度。留空启用动态分辨率。
+        * `--max_pixels`: 最大像素面积，动态分辨率时大于此值的图片会被缩小。
+* ERNIE-Image 专有参数
+    * `--tokenizer_path`: tokenizer 的路径，留空则自动从远程下载。
+
+我们构建了一个样例图像数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
+```
+
+我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/FLUX.md

@@ -195,7 +195,7 @@ FLUX 系列模型统一通过 [`examples/flux/model_training/train.py`](https://
 我们构建了一个样例图像数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/FLUX2.md

@@ -145,7 +145,7 @@ FLUX.2 系列模型统一通过 [`examples/flux2/model_training/train.py`](https
 我们构建了一个样例图像数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/JoyAI-Image.md

@@ -0,0 +1,154 @@
+# JoyAI-Image
+
+JoyAI-Image 是京东开源的统一多模态基础模型，支持图像理解、文生图生成和指令引导的图像编辑。
+
+## 安装
+
+在使用本项目进行模型推理和训练前，请先安装 DiffSynth-Studio。
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+更多关于安装的信息，请参考[安装依赖](../Pipeline_Usage/Setup.md)。
+
+## 快速开始
+
+运行以下代码可以快速加载 [jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 4G 显存即可运行。
+
+```python
+from diffsynth.pipelines.joyai_image import JoyAIImagePipeline, ModelConfig
+import torch
+from PIL import Image
+from modelscope import dataset_snapshot_download
+
+# Download dataset
+dataset_snapshot_download(
+    dataset_id="DiffSynth-Studio/diffsynth_example_dataset",
+    local_dir="data/diffsynth_example_dataset",
+    allow_file_pattern="joyai_image/JoyAI-Image-Edit/*"
+)
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+
+pipe = JoyAIImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="transformer/transformer.pth", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/model*.safetensors", **vram_config),
+        ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="vae/Wan2.1_VAE.pth", **vram_config),
+    ],
+    processor_config=ModelConfig(model_id="jd-opensource/JoyAI-Image-Edit", origin_file_pattern="JoyAI-Image-Und/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+# Use first sample from dataset
+dataset_base_path = "data/diffsynth_example_dataset/joyai_image/JoyAI-Image-Edit"
+prompt = "将裙子改为粉色"
+edit_image = Image.open(f"{dataset_base_path}/edit/image1.jpg").convert("RGB")
+
+output = pipe(
+    prompt=prompt,
+    edit_image=edit_image,
+    height=1024,
+    width=1024,
+    seed=0,
+    num_inference_steps=30,
+    cfg_scale=5.0,
+)
+
+output.save("output_joyai_edit_low_vram.png")
+```
+
+## 模型总览
+
+|模型 ID|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|
+|[jd-opensource/JoyAI-Image-Edit](https://modelscope.cn/models/jd-opensource/JoyAI-Image-Edit)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_inference/JoyAI-Image-Edit.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_inference_low_vram/JoyAI-Image-Edit.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/full/JoyAI-Image-Edit.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/validate_full/JoyAI-Image-Edit.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/lora/JoyAI-Image-Edit.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/joyai_image/model_training/validate_lora/JoyAI-Image-Edit.py)|
+
+## 模型推理
+
+模型通过 `JoyAIImagePipeline.from_pretrained` 加载，详见[加载模型](../Pipeline_Usage/Model_Inference.md#加载模型)。
+
+`JoyAIImagePipeline` 推理的输入参数包括：
+
+* `prompt`: 文本提示词，用于描述期望的图像编辑效果。
+* `negative_prompt`: 负向提示词，指定不希望出现在结果中的内容，默认为空字符串。
+* `cfg_scale`: 分类器自由引导的缩放系数，默认为 5.0。值越大，生成结果越贴近 prompt 描述。
+* `edit_image`: 待编辑的单张图像。
+* `denoising_strength`: 降噪强度，控制输入图像被重绘的程度，默认为 1.0。
+* `height`: 输出图像的高度，默认为 1024。需能被 16 整除。
+* `width`: 输出图像的宽度，默认为 1024。需能被 16 整除。
+* `seed`: 随机种子，用于控制生成的可复现性。设为 `None` 时使用随机种子。
+* `max_sequence_length`: 文本编码器处理的最大序列长度，默认为 4096。
+* `num_inference_steps`: 推理步数，默认为 30。步数越多，生成质量通常越好。
+* `tiled`: 是否启用分块处理，用于降低显存占用，默认为 False。
+* `tile_size`: 分块大小，默认为 (30, 52)。
+* `tile_stride`: 分块步幅，默认为 (15, 26)。
+* `shift`: 调度器的 shift 参数，用于控制 Flow Match 的调度曲线，默认为 4.0。
+* `progress_bar_cmd`: 进度条显示方式，默认为 tqdm。
+
+## 模型训练
+
+joyai_image 系列模型统一通过 `examples/joyai_image/model_training/train.py` 进行训练，脚本的参数包括：
+
+* 通用训练参数
+    * 数据集基础配置
+        * `--dataset_base_path`: 数据集的根目录。
+        * `--dataset_metadata_path`: 数据集的元数据文件路径。
+        * `--dataset_repeat`: 每个 epoch 中数据集重复的次数。
+        * `--dataset_num_workers`: 每个 Dataloader 的进程数量。
+        * `--data_file_keys`: 元数据中需要加载的字段名称，通常是图像或视频文件的路径，以 `,` 分隔。
+    * 模型加载配置
+        * `--model_paths`: 要加载的模型路径。JSON 格式。
+        * `--model_id_with_origin_paths`: 带原始路径的模型 ID。用逗号分隔。
+        * `--extra_inputs`: 模型 Pipeline 所需的额外输入参数，以 `,` 分隔。
+        * `--fp8_models`: 以 FP8 格式加载的模型，目前仅支持参数不被梯度更新的模型。
+    * 训练基础配置
+        * `--learning_rate`: 学习率。
+        * `--num_epochs`: 轮数（Epoch）。
+        * `--trainable_models`: 可训练的模型，例如 `dit`、`vae`、`text_encoder`。
+        * `--find_unused_parameters`: DDP 训练中是否存在未使用的参数。
+        * `--weight_decay`: 权重衰减大小。
+        * `--task`: 训练任务，默认为 `sft`。
+    * 输出配置
+        * `--output_path`: 模型保存路径。
+        * `--remove_prefix_in_ckpt`: 在模型文件的 state dict 中移除前缀。
+        * `--save_steps`: 保存模型的训练步数间隔。
+    * LoRA 配置
+        * `--lora_base_model`: LoRA 添加到哪个模型上。
+        * `--lora_target_modules`: LoRA 添加到哪些层上。
+        * `--lora_rank`: LoRA 的秩（Rank）。
+        * `--lora_checkpoint`: LoRA 检查点的路径。
+        * `--preset_lora_path`: 预置 LoRA 检查点路径，用于 LoRA 差分训练。
+        * `--preset_lora_model`: 预置 LoRA 融入的模型，例如 `dit`。
+    * 梯度配置
+        * `--use_gradient_checkpointing`: 是否启用 gradient checkpointing。
+        * `--use_gradient_checkpointing_offload`: 是否将 gradient checkpointing 卸载到内存中。
+        * `--gradient_accumulation_steps`: 梯度累积步数。
+    * 分辨率配置
+        * `--height`: 图像/视频的高度。留空启用动态分辨率。
+        * `--width`: 图像/视频的宽度。留空启用动态分辨率。
+        * `--max_pixels`: 最大像素面积，动态分辨率时大于此值的图片会被缩小。
+        * `--num_frames`: 视频的帧数（仅视频生成模型）。
+* JoyAI-Image 专有参数
+    * `--processor_path`: Processor 路径，用于处理文本和图像的编码器输入。
+    * `--initialize_model_on_cpu`: 是否在 CPU 上初始化模型，默认在加速设备上初始化。
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
+```
+
+关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/LTX-2.md

@@ -16,7 +16,7 @@ pip install -e .
 
 ## 快速开始
 
-运行以下代码可以快速加载 [Lightricks/LTX-2](https://www.modelscope.cn/models/Lightricks/LTX-2) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 8GB 显存即可运行。
+运行以下代码可以快速加载 [Lightricks/LTX-2.3](https://www.modelscope.cn/models/Lightricks/LTX-2.3) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 8GB 显存即可运行。
 
 ```python
 import torch
@@ -24,88 +24,36 @@ from diffsynth.pipelines.ltx2_audio_video import LTX2AudioVideoPipeline, ModelCo
 from diffsynth.utils.data.media_io_ltx2 import write_video_audio_ltx2
 
 vram_config = {
-    "offload_dtype": torch.float8_e5m2,
+    "offload_dtype": torch.bfloat16,
     "offload_device": "cpu",
-    "onload_dtype": torch.float8_e5m2,
-    "onload_device": "cpu",
-    "preparing_dtype": torch.float8_e5m2,
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cuda",
+    "preparing_dtype": torch.bfloat16,
     "preparing_device": "cuda",
     "computation_dtype": torch.bfloat16,
     "computation_device": "cuda",
 }
-"""
-Offical model repo: https://www.modelscope.cn/models/Lightricks/LTX-2
-Repackaged model repo: https://www.modelscope.cn/models/DiffSynth-Studio/LTX-2-Repackage
-For base models of LTX-2, offical checkpoint (with model config ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-dev.safetensors"))
-and repackaged checkpoints (with model config ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="*.safetensors")) are both supported.
-We have repackeged the official checkpoints in DiffSynth-Studio/LTX-2-Repackage repo to support separate loading of different submodules,
-and avoid redundant memory usage when users only want to use part of the model.
-"""
-# use the repackaged modelconfig from "DiffSynth-Studio/LTX-2-Repackage" to avoid redundant model loading
 pipe = LTX2AudioVideoPipeline.from_pretrained(
     torch_dtype=torch.bfloat16,
     device="cuda",
     model_configs=[
         ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized", origin_file_pattern="model-*.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="transformer.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="text_encoder_post_modules.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="video_vae_decoder.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="audio_vae_decoder.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="audio_vocoder.safetensors", **vram_config),
-        ModelConfig(model_id="DiffSynth-Studio/LTX-2-Repackage", origin_file_pattern="video_vae_encoder.safetensors", **vram_config),
-        ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-spatial-upscaler-x2-1.0.safetensors", **vram_config),
+        ModelConfig(model_id="Lightricks/LTX-2.3", origin_file_pattern="ltx-2.3-22b-dev.safetensors", **vram_config),
+        ModelConfig(model_id="Lightricks/LTX-2.3", origin_file_pattern="ltx-2.3-spatial-upscaler-x2-1.0.safetensors", **vram_config),
     ],
     tokenizer_config=ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized"),
-    stage2_lora_config=ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-distilled-lora-384.safetensors"),
-    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+    stage2_lora_config=ModelConfig(model_id="Lightricks/LTX-2.3", origin_file_pattern="ltx-2.3-22b-distilled-lora-384.safetensors"),
 )
-
-# use the following modelconfig if you want to initialize model from offical checkpoints from "Lightricks/LTX-2"
-# pipe = LTX2AudioVideoPipeline.from_pretrained(
-#     torch_dtype=torch.bfloat16,
-#     device="cuda",
-#     model_configs=[
-#         ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized", origin_file_pattern="model-*.safetensors", **vram_config),
-#         ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-dev.safetensors", **vram_config),
-#         ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-spatial-upscaler-x2-1.0.safetensors", **vram_config),
-#     ],
-#     tokenizer_config=ModelConfig(model_id="google/gemma-3-12b-it-qat-q4_0-unquantized"),
-#     stage2_lora_config=ModelConfig(model_id="Lightricks/LTX-2", origin_file_pattern="ltx-2-19b-distilled-lora-384.safetensors"),
-#     vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
-# )
-
-prompt = "A girl is very happy, she is speaking: \"I enjoy working with Diffsynth-Studio, it's a perfect framework.\""
-negative_prompt = (
-    "blurry, out of focus, overexposed, underexposed, low contrast, washed out colors, excessive noise, "
-    "grainy texture, poor lighting, flickering, motion blur, distorted proportions, unnatural skin tones, "
-    "deformed facial features, asymmetrical face, missing facial features, extra limbs, disfigured hands, "
-    "wrong hand count, artifacts around text, inconsistent perspective, camera shake, incorrect depth of "
-    "field, background too sharp, background clutter, distracting reflections, harsh shadows, inconsistent "
-    "lighting direction, color banding, cartoonish rendering, 3D CGI look, unrealistic materials, uncanny "
-    "valley effect, incorrect ethnicity, wrong gender, exaggerated expressions, wrong gaze direction, "
-    "mismatched lip sync, silent or muted audio, distorted voice, robotic voice, echo, background noise, "
-    "off-sync audio, incorrect dialogue, added dialogue, repetitive speech, jittery movement, awkward "
-    "pauses, incorrect timing, unnatural transitions, inconsistent framing, tilted camera, flat lighting, "
-    "inconsistent tone, cinematic oversaturation, stylized filters, or AI artifacts."
-)
-height, width, num_frames = 512 * 2, 768 * 2, 121
+prompt = "Two cute orange cats, wearing boxing gloves, stand in a boxing ring and fight each other. They are punching each other fast and yelling: 'I will win!'"
+negative_prompt = pipe.default_negative_prompt["LTX-2.3"]
 video, audio = pipe(
     prompt=prompt,
     negative_prompt=negative_prompt,
     seed=43,
-    height=height,
-    width=width,
-    num_frames=num_frames,
-    tiled=True,
-    use_two_stage_pipeline=True,
-)
-write_video_audio_ltx2(
-    video=video,
-    audio=audio,
-    output_path='ltx2_twostage.mp4',
-    fps=24,
-    audio_sample_rate=24000,
+    height=1024, width=1536, num_frames=121,
+    tiled=True, use_two_stage_pipeline=True,
 )
+write_video_audio_ltx2(video=video, audio=audio, output_path='video.mp4', fps=24, audio_sample_rate=pipe.audio_vocoder.output_sampling_rate)
 ```
 
 ## 模型总览
@@ -217,7 +165,7 @@ LTX-2 系列模型统一通过 [`examples/ltx2/model_training/train.py`](https:/
 我们构建了一个样例视频数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_video_dataset --local_dir ./data/example_video_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/Qwen-Image.md

@@ -199,7 +199,7 @@ Qwen-Image 系列模型统一通过 [`examples/qwen_image/model_training/train.p
 我们构建了一个样例图像数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文“模型总览”中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/Stable-Diffusion-XL.md

@@ -0,0 +1,141 @@
+# Stable Diffusion XL
+
+Stable Diffusion XL (SDXL) 是由 Stability AI 开发的开源扩散式文本到图像生成模型，支持 1024x1024 分辨率的高质量文本到图像生成，采用双文本编码器（CLIP-L + CLIP-bigG）架构。
+
+## 安装
+
+在使用本项目进行模型推理和训练前，请先安装 DiffSynth-Studio。
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+更多关于安装的信息，请参考[安装依赖](../Pipeline_Usage/Setup.md)。
+
+## 快速开始
+
+运行以下代码可以快速加载 [stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 6GB 显存即可运行。
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion_xl import StableDiffusionXLPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="text_encoder_2/model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer/"),
+    tokenizer_2_config=ModelConfig(model_id="stabilityai/stable-diffusion-xl-base-1.0", origin_file_pattern="tokenizer_2/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars",
+    negative_prompt="",
+    cfg_scale=5.0,
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+## 模型总览
+
+|模型 ID|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|
+|[stabilityai/stable-diffusion-xl-base-1.0](https://www.modelscope.cn/models/stabilityai/stable-diffusion-xl-base-1.0)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_inference/stable-diffusion-xl-base-1.0.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_inference_low_vram/stable-diffusion-xl-base-1.0.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/full/stable-diffusion-xl-base-1.0.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/validate_full/stable-diffusion-xl-base-1.0.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/lora/stable-diffusion-xl-base-1.0.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/validate_lora/stable-diffusion-xl-base-1.0.py)|
+
+## 模型推理
+
+模型通过 `StableDiffusionXLPipeline.from_pretrained` 加载，详见[加载模型](../Pipeline_Usage/Model_Inference.md#加载模型)。
+
+`StableDiffusionXLPipeline` 的推理输入参数包括：
+
+* `prompt`: 文本提示词。
+* `negative_prompt`: 负面提示词，默认为空字符串。
+* `cfg_scale`: Classifier-Free Guidance 缩放系数，默认 5.0。
+* `height`: 输出图像高度，默认 1024。
+* `width`: 输出图像宽度，默认 1024。
+* `seed`: 随机种子，默认不设置时使用随机种子。
+* `rand_device`: 噪声生成设备，默认 "cpu"。
+* `num_inference_steps`: 推理步数，默认 50。
+* `guidance_rescale`: Guidance rescale 系数，默认 0.0。
+* `progress_bar_cmd`: 进度条回调函数。
+
+> `StableDiffusionXLPipeline` 需要双 tokenizer 配置（`tokenizer_config` 和 `tokenizer_2_config`），分别对应 CLIP-L 和 CLIP-bigG 文本编码器。
+
+## 模型训练
+
+stable_diffusion_xl 系列模型通过 `examples/stable_diffusion_xl/model_training/train.py` 进行训练，脚本的参数包括：
+
+* 通用训练参数
+    * 数据集基础配置
+        * `--dataset_base_path`: 数据集的根目录。
+        * `--dataset_metadata_path`: 数据集的元数据文件路径。
+        * `--dataset_repeat`: 每个 epoch 中数据集重复的次数。
+        * `--dataset_num_workers`: 每个 Dataloader 的进程数量。
+        * `--data_file_keys`: 元数据中需要加载的字段名称，通常是图像或视频文件的路径，以 `,` 分隔。
+    * 模型加载配置
+        * `--model_paths`: 要加载的模型路径。JSON 格式。
+        * `--model_id_with_origin_paths`: 带原始路径的模型 ID。用逗号分隔。
+        * `--extra_inputs`: 模型 Pipeline 所需的额外输入参数，以 `,` 分隔。
+        * `--fp8_models`: 以 FP8 格式加载的模型，目前仅支持参数不被梯度更新的模型。
+    * 训练基础配置
+        * `--learning_rate`: 学习率。
+        * `--num_epochs`: 轮数（Epoch）。
+        * `--trainable_models`: 可训练的模型，例如 `dit`、`vae`、`text_encoder`。
+        * `--find_unused_parameters`: DDP 训练中是否存在未使用的参数。
+        * `--weight_decay`: 权重衰减大小。
+        * `--task`: 训练任务，默认为 `sft`。
+    * 输出配置
+        * `--output_path`: 模型保存路径。
+        * `--remove_prefix_in_ckpt`: 在模型文件的 state dict 中移除前缀。
+        * `--save_steps`: 保存模型的训练步数间隔。
+    * LoRA 配置
+        * `--lora_base_model`: LoRA 添加到哪个模型上。
+        * `--lora_target_modules`: LoRA 添加到哪些层上。
+        * `--lora_rank`: LoRA 的秩（Rank）。
+        * `--lora_checkpoint`: LoRA 检查点的路径。
+        * `--preset_lora_path`: 预置 LoRA 检查点路径，用于 LoRA 差分训练。
+        * `--preset_lora_model`: 预置 LoRA 融入的模型，例如 `dit`。
+    * 梯度配置
+        * `--use_gradient_checkpointing`: 是否启用 gradient checkpointing。
+        * `--use_gradient_checkpointing_offload`: 是否将 gradient checkpointing 卸载到内存中。
+        * `--gradient_accumulation_steps`: 梯度累积步数。
+    * 分辨率配置
+        * `--height`: 图像/视频的高度。留空启用动态分辨率。
+        * `--width`: 图像/视频的宽度。留空启用动态分辨率。
+        * `--max_pixels`: 最大像素面积，动态分辨率时大于此值的图片会被缩小。
+        * `--num_frames`: 视频的帧数（仅视频生成模型）。
+* Stable Diffusion XL 专有参数
+    * `--tokenizer_path`: 第一个 Tokenizer 路径。
+    * `--tokenizer_2_path`: 第二个 Tokenizer 路径，默认为 `stabilityai/stable-diffusion-xl-base-1.0:tokenizer_2/`。
+
+样例数据集下载：
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "stable_diffusion_xl/*" --local_dir ./data/diffsynth_example_dataset
+```
+
+[stable-diffusion-xl-base-1.0 训练脚本](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion_xl/model_training/lora/stable-diffusion-xl-base-1.0.sh)
+
+我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/Stable-Diffusion.md

@@ -0,0 +1,138 @@
+# Stable Diffusion
+
+Stable Diffusion 是由 Stability AI 开发的开源扩散式文本到图像生成模型，支持 512x512 分辨率的文本到图像生成。
+
+## 安装
+
+在使用本项目进行模型推理和训练前，请先安装 DiffSynth-Studio。
+
+```shell
+git clone https://github.com/modelscope/DiffSynth-Studio.git
+cd DiffSynth-Studio
+pip install -e .
+```
+
+更多关于安装的信息，请参考[安装依赖](../Pipeline_Usage/Setup.md)。
+
+## 快速开始
+
+运行以下代码可以快速加载 [AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5) 模型并进行推理。显存管理已启动，框架会自动根据剩余显存控制模型参数的加载，最低 2GB 显存即可运行。
+
+```python
+import torch
+from diffsynth.core import ModelConfig
+from diffsynth.pipelines.stable_diffusion import StableDiffusionPipeline
+
+vram_config = {
+    "offload_dtype": torch.float32,
+    "offload_device": "cpu",
+    "onload_dtype": torch.float32,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.float32,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.float32,
+    "computation_device": "cuda",
+}
+pipe = StableDiffusionPipeline.from_pretrained(
+    torch_dtype=torch.float32,
+    model_configs=[
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="unet/diffusion_pytorch_model.safetensors", **vram_config),
+        ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="AI-ModelScope/stable-diffusion-v1-5", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="a photo of an astronaut riding a horse on mars, high quality, detailed",
+    negative_prompt="blurry, low quality, deformed",
+    cfg_scale=7.5,
+    height=512,
+    width=512,
+    seed=42,
+    rand_device="cuda",
+    num_inference_steps=50,
+)
+image.save("image.jpg")
+```
+
+## 模型总览
+
+|模型 ID|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|
+|[AI-ModelScope/stable-diffusion-v1-5](https://www.modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_inference/stable-diffusion-v1-5.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_inference_low_vram/stable-diffusion-v1-5.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/full/stable-diffusion-v1-5.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/validate_full/stable-diffusion-v1-5.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/lora/stable-diffusion-v1-5.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/validate_lora/stable-diffusion-v1-5.py)|
+
+## 模型推理
+
+模型通过 `StableDiffusionPipeline.from_pretrained` 加载，详见[加载模型](../Pipeline_Usage/Model_Inference.md#加载模型)。
+
+`StableDiffusionPipeline` 的推理输入参数包括：
+
+* `prompt`: 文本提示词。
+* `negative_prompt`: 负面提示词，默认为空字符串。
+* `cfg_scale`: Classifier-Free Guidance 缩放系数，默认 7.5。
+* `height`: 输出图像高度，默认 512。
+* `width`: 输出图像宽度，默认 512。
+* `seed`: 随机种子，默认不设置时使用随机种子。
+* `rand_device`: 噪声生成设备，默认 "cpu"。
+* `num_inference_steps`: 推理步数，默认 50。
+* `eta`: DDIM 调度器的 eta 参数，默认 0.0。
+* `guidance_rescale`: Guidance rescale 系数，默认 0.0。
+* `progress_bar_cmd`: 进度条回调函数。
+
+## 模型训练
+
+stable_diffusion 系列模型通过 `examples/stable_diffusion/model_training/train.py` 进行训练，脚本的参数包括：
+
+* 通用训练参数
+    * 数据集基础配置
+        * `--dataset_base_path`: 数据集的根目录。
+        * `--dataset_metadata_path`: 数据集的元数据文件路径。
+        * `--dataset_repeat`: 每个 epoch 中数据集重复的次数。
+        * `--dataset_num_workers`: 每个 Dataloader 的进程数量。
+        * `--data_file_keys`: 元数据中需要加载的字段名称，通常是图像或视频文件的路径，以 `,` 分隔。
+    * 模型加载配置
+        * `--model_paths`: 要加载的模型路径。JSON 格式。
+        * `--model_id_with_origin_paths`: 带原始路径的模型 ID。用逗号分隔。
+        * `--extra_inputs`: 模型 Pipeline 所需的额外输入参数，以 `,` 分隔。
+        * `--fp8_models`: 以 FP8 格式加载的模型，目前仅支持参数不被梯度更新的模型。
+    * 训练基础配置
+        * `--learning_rate`: 学习率。
+        * `--num_epochs`: 轮数（Epoch）。
+        * `--trainable_models`: 可训练的模型，例如 `dit`、`vae`、`text_encoder`。
+        * `--find_unused_parameters`: DDP 训练中是否存在未使用的参数。
+        * `--weight_decay`: 权重衰减大小。
+        * `--task`: 训练任务，默认为 `sft`。
+    * 输出配置
+        * `--output_path`: 模型保存路径。
+        * `--remove_prefix_in_ckpt`: 在模型文件的 state dict 中移除前缀。
+        * `--save_steps`: 保存模型的训练步数间隔。
+    * LoRA 配置
+        * `--lora_base_model`: LoRA 添加到哪个模型上。
+        * `--lora_target_modules`: LoRA 添加到哪些层上。
+        * `--lora_rank`: LoRA 的秩（Rank）。
+        * `--lora_checkpoint`: LoRA 检查点的路径。
+        * `--preset_lora_path`: 预置 LoRA 检查点路径，用于 LoRA 差分训练。
+        * `--preset_lora_model`: 预置 LoRA 融入的模型，例如 `dit`。
+    * 梯度配置
+        * `--use_gradient_checkpointing`: 是否启用 gradient checkpointing。
+        * `--use_gradient_checkpointing_offload`: 是否将 gradient checkpointing 卸载到内存中。
+        * `--gradient_accumulation_steps`: 梯度累积步数。
+    * 分辨率配置
+        * `--height`: 图像/视频的高度。留空启用动态分辨率。
+        * `--width`: 图像/视频的宽度。留空启用动态分辨率。
+        * `--max_pixels`: 最大像素面积，动态分辨率时大于此值的图片会被缩小。
+        * `--num_frames`: 视频的帧数（仅视频生成模型）。
+* Stable Diffusion 专有参数
+    * `--tokenizer_path`: Tokenizer 路径，默认为 `AI-ModelScope/stable-diffusion-v1-5:tokenizer/`。
+
+样例数据集下载：
+
+```shell
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "stable_diffusion/*" --local_dir ./data/diffsynth_example_dataset
+```
+
+[stable-diffusion-v1-5 训练脚本](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/stable_diffusion/model_training/lora/stable-diffusion-v1-5.sh)
+
+我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/Wan.md

@@ -105,41 +105,43 @@ graph LR;
 
 </details>
 
-|模型 ID|额外参数|推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
-|-|-|-|-|-|-|-|
-|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
-|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
-|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
-|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
-|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
-|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
-|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
-|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
-|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
-|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
-|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
-|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
-|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
-|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
-|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
-|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
-|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
-|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
-|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
-|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
-|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
-|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
-|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
-|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
-|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
-|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
-|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
-| [openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-360p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-360P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-360P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py) |
-| [openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p) | `input_image` | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-720p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-720P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-720P-I2AV.sh) | [code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py) |
+|模型 ID|额外参数|推理|低显存推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
+|-|-|-|-|-|-|-|-|
+|[Wan-AI/Wan2.1-T2V-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-1.3B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-1.3B.py)|
+|[Wan-AI/Wan2.1-T2V-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-T2V-14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-T2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-T2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-T2V-14B.py)|
+|[Wan-AI/Wan2.1-I2V-14B-480P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-480P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-480P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-480P.py)|
+|[Wan-AI/Wan2.1-I2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-I2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-I2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-I2V-14B-720P.py)|
+|[Wan-AI/Wan2.1-FLF2V-14B-720P](https://modelscope.cn/models/Wan-AI/Wan2.1-FLF2V-14B-720P)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-FLF2V-14B-720P.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-FLF2V-14B-720P.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-FLF2V-14B-720P.py)|
+|[iic/VACE-Wan2.1-1.3B-Preview](https://modelscope.cn/models/iic/VACE-Wan2.1-1.3B-Preview)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B-Preview.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B-Preview.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B-Preview.py)|
+|[Wan-AI/Wan2.1-VACE-1.3B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-1.3B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-1.3B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-1.3B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-1.3B.py)|
+|[Wan-AI/Wan2.1-VACE-14B](https://modelscope.cn/models/Wan-AI/Wan2.1-VACE-14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-VACE-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-VACE-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-VACE-14B.py)|
+|[PAI/Wan2.1-Fun-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-1.3B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-InP.py)|
+|[PAI/Wan2.1-Fun-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-14B-Control)|`control_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-InP](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-InP.py)|
+|[PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-1.3B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-1.3B-Control-Camera.py)|
+|[PAI/Wan2.1-Fun-V1.1-14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.1-Fun-V1.1-14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-Fun-V1.1-14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-Fun-V1.1-14B-Control-Camera.py)|
+|[DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1](https://modelscope.cn/models/DiffSynth-Studio/Wan2.1-1.3b-speedcontrol-v1)|`motion_bucket_id`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.1-1.3b-speedcontrol-v1.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.1-1.3b-speedcontrol-v1.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.1-1.3b-speedcontrol-v1.py)|
+|[krea/krea-realtime-video](https://www.modelscope.cn/models/krea/krea-realtime-video)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/krea-realtime-video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/krea-realtime-video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/krea-realtime-video.py)|
+|[meituan-longcat/LongCat-Video](https://www.modelscope.cn/models/meituan-longcat/LongCat-Video)|`longcat_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/LongCat-Video.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/LongCat-Video.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/LongCat-Video.py)|
+|[ByteDance/Video-As-Prompt-Wan2.1-14B](https://modelscope.cn/models/ByteDance/Video-As-Prompt-Wan2.1-14B)|`vap_video`, `vap_prompt`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Video-As-Prompt-Wan2.1-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Video-As-Prompt-Wan2.1-14B.py)|
+|[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
+|[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
+|[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|
+|[Wan-AI/Wan2.2-Animate-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-Animate-14B)|`input_image`, `animate_pose_video`, `animate_face_video`, `animate_inpaint_video`, `animate_mask_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Animate-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Animate-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Animate-14B.py)|
+|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-S2V-14B_multi_clips.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-S2V-14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-S2V-14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-S2V-14B.py)|
+|[PAI/Wan2.2-VACE-Fun-A14B](https://www.modelscope.cn/models/PAI/Wan2.2-VACE-Fun-A14B)|`vace_control_video`, `vace_reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-VACE-Fun-A14B.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-VACE-Fun-A14B.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-VACE-Fun-A14B.py)|
+|[PAI/Wan2.2-Fun-A14B-InP](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-InP)|`input_image`, `end_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-InP.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-InP.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-InP.py)|
+|[PAI/Wan2.2-Fun-A14B-Control](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control)|`control_video`, `reference_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control.py)|
+|[PAI/Wan2.2-Fun-A14B-Control-Camera](https://modelscope.cn/models/PAI/Wan2.2-Fun-A14B-Control-Camera)|`control_camera_video`, `input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/Wan2.2-Fun-A14B-Control-Camera.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/Wan2.2-Fun-A14B-Control-Camera.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/Wan2.2-Fun-A14B-Control-Camera.py)|
+|[openmoss/MOVA-360p](https://modelscope.cn/models/openmoss/MOVA-360p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-360p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-360P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-360p-I2AV.py)|
+|[openmoss/MOVA-720p](https://modelscope.cn/models/openmoss/MOVA-720p)|`input_image`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_inference_low_vram/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/full/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_full/MOVA-720p-I2AV.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/lora/MOVA-720P-I2AV.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/mova/model_training/validate_lora/MOVA-720p-I2AV.py)|
+|[Wan-AI/WanToDance-14B (global model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-global.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-global.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-global.py)|
+|[Wan-AI/WanToDance-14B (local model)](https://modelscope.cn/models/Wan-AI/WanToDance-14B)|`wantodance_music_path`, `wantodance_reference_image`, `wantodance_fps`, `wantodance_keyframes`, `wantodance_keyframes_mask`|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_inference_low_vram/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/full/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_full/WanToDance-14B-local.py)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/lora/WanToDance-14B-local.sh)|[code](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/validate_lora/WanToDance-14B-local.py)|
 
 * FP8 精度训练：[doc](../Training/FP8_Precision.md)、[code](https://github.com/modelscope/DiffSynth-Studio/tree/main/examples/wanvideo/model_training/special/fp8_training/)
 * 两阶段拆分训练：[doc](../Training/Split_Training.md)、[code](https://github.com/modelscope/DiffSynth-Studio/tree/main/examples/wanvideo/model_training/special/split_training/)
@@ -204,6 +206,50 @@ DeepSpeed ZeRO 3 训练：Wan 系列模型支持 DeepSpeed ZeRO 3 训练，将
 
 如果显存不足，请开启[显存管理](../Pipeline_Usage/VRAM_management.md)，我们在示例代码中提供了每个模型推荐的低显存配置，详见前文"模型总览"中的表格。
 
+### 多卡并行加速
+
+如需开启多卡并行加速，请先安装 `flash_attn` 与 `xfuser`：
+
+```shell
+pip install flash-attn --no-build-isolation
+pip install xfuser
+```
+
+对代码进行如下修改（[样例代码](https://github.com/modelscope/DiffSynth-Studio/tree/main/examples/wanvideo/acceleration/unified_sequence_parallel.py)）：
+
+```diff
+import torch
+from PIL import Image
+from diffsynth.utils.data import save_video, VideoData
+from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
++ import torch.distributed as dist
+
+pipe = WanVideoPipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
++   use_usp=True,
+    model_configs=[
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="models_t5_umt5-xxl-enc-bf16.pth"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="Wan2.1_VAE.pth"),
+    ],
+    tokenizer_config=ModelConfig(model_id="Wan-AI/Wan2.1-T2V-1.3B", origin_file_pattern="google/umt5-xxl/"),
+)
+video = pipe(
+    prompt="一名宇航员身穿太空服，面朝镜头骑着一匹机械马在火星表面驰骋。红色的荒凉地表延伸至远方，点缀着巨大的陨石坑和奇特的岩石结构。机械马的步伐稳健，扬起微弱的尘埃，展现出未来科技与原始探索的完美结合。宇航员手持操控装置，目光坚定，仿佛正在开辟人类的新疆域。背景是深邃的宇宙和蔚蓝的地球，画面既科幻又充满希望，让人不禁畅想未来的星际生活。",
+    negative_prompt="色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走",
+    seed=0, tiled=True,
+)
++ if dist.get_rank() == 0:
++   save_video(video, "video1.mp4", fps=15, quality=5)
+```
+
+运行多卡并行推理时，请使用 `torchrun` 运行，其中 `--nproc_per_node` 为 GPU 数量：
+
+```shell
+torchrun --nproc_per_node=8 examples/wanvideo/acceleration/unified_sequence_parallel.py
+```
+
 ## 模型训练
 
 Wan 系列模型统一通过 [`examples/wanvideo/model_training/train.py`](https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/model_training/train.py) 进行训练，脚本的参数包括：
@@ -254,7 +300,7 @@ Wan 系列模型统一通过 [`examples/wanvideo/model_training/train.py`](https
 我们构建了一个样例视频数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_video_dataset --local_dir ./data/example_video_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Model_Details/Z-Image.md

@@ -134,7 +134,7 @@ Z-Image 系列模型统一通过 [`examples/z_image/model_training/train.py`](ht
 我们构建了一个样例图像数据集，以方便您进行测试，通过以下命令可以下载这个数据集：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 我们为每个模型编写了推荐的训练脚本，请参考前文"模型总览"中的表格。关于如何编写模型训练脚本，请参考[模型训练](../Pipeline_Usage/Model_Training.md)；更多高阶训练算法，请参考[训练框架详解](https://github.com/modelscope/DiffSynth-Studio/tree/main/docs/zh/Training/)。

docs/zh/Pipeline_Usage/Accelerated_Inference.md

@@ -0,0 +1,84 @@
+# 推理加速
+
+扩散模型的去噪过程通常耗时较长。为提升推理速度，可采用多种加速技术，包含多卡并行推理、计算图编译等无损加速方案，以及 Cache、量化等有损加速方案。
+
+当前扩散模型大多基于 Diffusion Transformer 构建，高效注意力机制同样是常用的加速手段。DiffSynth-Studio 目前已支持部分无损加速推理功能。本节重点从多卡并行推理和计算图编译两个维度介绍加速方法。
+
+## 高效注意力机制
+注意力机制的加速细节请参考 [注意力机制实现](../API_Reference/core/attention.md)。
+
+## 多卡并行推理
+DiffSynth-Studio 采用统一序列并行的多卡推理方案。在 DiT 中将 token 序列拆分至多张显卡进行并行处理。底层基于 [xDiT](https://github.com/xdit-project/xDiT) 实现。需要注意，统一序列并行会引入额外通信开销，实际加速比通常低于显卡数量。
+
+目前 DiffSynth-Studio 已支持 [Wan](../Model_Details/Wan.md) 和 [MOVA](../Model_Details/Wan.md) 模型的统一序列并行加速。
+
+首先安装 `xDiT` 依赖。
+```bash
+pip install "xfuser[flash-attn]>=0.4.3"
+```
+
+然后使用 `torchrun` 启动多卡推理。
+```bash
+torchrun --standalone --nproc_per_node=8 examples/wanvideo/acceleration/unified_sequence_parallel.py
+```
+
+构建 pipeline 时配置 `usp=True` 即可实现 USP 并行推理。代码示例如下。
+```python
+import torch
+from PIL import Image
+from diffsynth.utils.data import save_video
+from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
+import torch.distributed as dist
+
+pipe = WanVideoPipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    use_usp=True,
+    model_configs=[
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="models_t5_umt5-xxl-enc-bf16.pth"),
+        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-14B", origin_file_pattern="Wan2.1_VAE.pth"),
+    ],
+    tokenizer_config=ModelConfig(model_id="Wan-AI/Wan2.1-T2V-1.3B", origin_file_pattern="google/umt5-xxl/"),
+)
+
+# Text-to-video
+video = pipe(
+    prompt="一名宇航员身穿太空服，面朝镜头骑着一匹机械马在火星表面驰骋。红色的荒凉地表延伸至远方，点缀着巨大的陨石坑和奇特的岩石结构。机械马的步伐稳健，扬起微弱的尘埃，展现出未来科技与原始探索的完美结合。宇航员手持操控装置，目光坚定，仿佛正在开辟人类的新疆域。背景是深邃的宇宙和蔚蓝的地球，画面既科幻又充满希望，让人不禁畅想未来的星际生活。",
+    negative_prompt="色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走",
+    seed=0, tiled=True,
+)
+if dist.get_rank() == 0:
+    save_video(video, "video1.mp4", fps=15, quality=5)
+```
+
+## 计算图编译
+PyTorch 2.0 提供了自动计算图编译接口 [torch.compile](https://docs.pytorch.org/tutorials/intermediate/torch_compile_tutorial.html)，能够将 PyTorch 代码即时编译为优化内核，从而提升运行速度。由于扩散模型的推理耗时集中在 DiT 的多步去噪阶段，且 DiT 主要由基础模块堆叠而成，为缩短编译时间，DiffSynth 的 compile 功能采用仅针对基础 Transformer 模块的 [区域编译](https://docs.pytorch.org/tutorials/recipes/regional_compilation.html) 策略。
+
+### Compile 使用示例
+相比常规推理，只需在调用 pipeline 前执行 `pipe.compile_pipeline()` 即可开启编译加速。具体函数定义请参阅[源代码](https://github.com/modelscope/DiffSynth-Studio/blob/166e6d2d38764209f66a74dd0fe468226536ad0f/diffsynth/diffusion/base_pipeline.py#L342)。
+
+`compile_pipeline` 的输入参数主要包含两类。
+
+第一类是编译模型参数 `compile_models`。以 Qwen-Image Pipeline 为例，若仅编译 DiT 模型，保持该参数为空即可。若需额外编译 VAE 等模型，可传入 `compile_models=["vae", "dit"]`。除 DiT 外，其余模型均采用整体编译策略，即把模型的 forward 函数完整编译为计算图。
+
+第二类是编译策略参数。涵盖 `mode`, `dynamic`, `fullgraph` 及其他自定义选项。这些参数会直接传递给 `torch.compile` 接口。若未深入了解这些参数的具体机制，建议保持默认设置。
+
+- `mode` 指定编译模式，包含 `"default"`, `"reduce-overhead"`, `"max-autotune"` 和 `"max-autotune-no-cudagraphs"`。由于 cudagraph 对计算图要求较为严格（例如可能需要配合 `torch.compiler.cudagraph_mark_step_begin()` 使用），`"reduce-overhead"` 和 `"max-autotune"` 模式可能编译失败。
+- `dynamic` 决定是否启用动态形状。对于多数生成模型，修改 prompt、开启 CFG 或调整分辨率都会改变计算图的输入张量形状。设置为 `dynamic=True` 会增加首次运行的编译时长，但支持动态形状，形状改变时无需重编译。设置为 `dynamic=False` 时首次编译较快，但任何改变输入形状的操作都会触发重新编译。对大部分场景，建议设定为 `dynamic=True`。
+- `fullgraph` 设为 `True` 时，底层会尝试将目标模型编译为单一计算图，若失败则报错。设为 `False` 时，底层会在无法连接处设置断点，将模型编译为多个独立计算图。开发者可开启 `True` 来优化编译性能，普通用户建议仅使用 `False`。
+- 其他参数配置请查阅 [api 文档](https://docs.pytorch.org/docs/stable/generated/torch.compile.html)。
+
+### Compile 功能开发者文档
+若需为新接入的 pipeline 提供 compile 支持，应在 pipeline 中配置 `compilable_models` 属性以指定默认编译模型。针对该 pipeline 的 DiT 模型类，还需配置 `_repeated_blocks` 以指定参与区域编译的基础模块类型。
+
+以 Qwen-Image 为例，其 pipeline 配置如下。
+```python
+self.compilable_models = ["dit"]
+```
+
+其 DiT 配置如下。
+```python
+class QwenImageDiT(torch.nn.Module):
+    _repeated_blocks = ["QwenImageTransformerBlock"]
+```

docs/zh/Pipeline_Usage/Model_Training.md

@@ -69,28 +69,16 @@ image_2.jpg,"a cat"
 
 <details>
 
-<summary>样例图像数据集</summary>
+<summary>样例数据集</summary>
 
 > ```shell
-> modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+> modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 > ```
 > 
 > 适用于 Qwen-Image、FLUX 等图像生成模型的训练。
 
 </details>
 
-<details>
-
-<summary>样例视频数据集</summary>
-
-> ```shell
-> modelscope download --dataset DiffSynth-Studio/example_video_dataset --local_dir ./data/example_video_dataset
-> ```
-> 
-> 适用于 Wan 等视频生成模型的训练。
-
-</details>
-
 ## 加载模型
 
 类似于[推理时的模型加载](../Pipeline_Usage/Model_Inference.md#加载模型)，我们支持多种方式配置模型路径，两种方式是可以混用的。
@@ -243,3 +231,116 @@ accelerate launch --config_file examples/qwen_image/model_training/full/accelera
 * 少数模型包含冗余参数，例如 Qwen-Image 的 DiT 部分最后一层的文本编码部分，在训练这些模型时，需设置 `--find_unused_parameters` 避免在多 GPU 训练中报错。出于对开源社区模型兼容性的考虑，我们不打算删除这些冗余参数。
 * Diffusion 模型的损失函数值与实际效果的关系不大，因此我们在训练过程中不会记录损失函数值。我们建议把 `--num_epochs` 设置为足够大的数值，边训边测，直至效果收敛后手动关闭训练程序。
 * `--use_gradient_checkpointing` 通常是开启的，除非 GPU 显存足够；`--use_gradient_checkpointing_offload` 则按需开启，详见 [`diffsynth.core.gradient`](../API_Reference/core/gradient.md)。
+
+## 低显存训练
+如果想在低显存显卡上完成 LoRA 模型训练，可以同时采用 [两阶段拆分训练](../Training/Split_Training.md) 和 `deepspeed_zero3_offload` 训练。 首先，将前处理过程拆分到第一阶段，将计算结果存储到硬盘中。其次，在第二阶段从硬盘中读取这些结果并进行去噪模型的训练，训练通过采用 `deepspeed_zero3_offload`，将训练参数和优化器状态 offload 到 cpu 或者 disk 上。我们为部分模型提供了样例，主要是通过 `--config_file` 指定 `deepspeed` 配置。
+
+需要注意的是，`deepspeed_zero3_offload` 模式与 `pytorch` 原生的梯度检查点机制不兼容，我们为此对 `deepspeed` 的`checkpointing` 接口做了适配。用户需要在 `deepspeed` 配置中填写 `activation_checkpointing` 字段以启用梯度检查点。
+
+以下为 Qwen-Image 模型的低显存模型训练脚本：
+```shell
+accelerate launch examples/qwen_image/model_training/train.py \
+  --dataset_base_path data/example_image_dataset \
+  --dataset_metadata_path data/example_image_dataset/metadata.csv \
+  --max_pixels 1048576 \
+  --dataset_repeat 1 \
+  --model_id_with_origin_paths "Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \
+  --learning_rate 1e-4 \
+  --num_epochs 5 \
+  --remove_prefix_in_ckpt "pipe.dit." \
+  --output_path "./models/train/Qwen-Image_lora-splited-cache" \
+  --lora_base_model "dit" \
+  --lora_target_modules "to_q,to_k,to_v,add_q_proj,add_k_proj,add_v_proj,to_out.0,to_add_out,img_mlp.net.2,img_mod.1,txt_mlp.net.2,txt_mod.1" \
+  --lora_rank 32 \
+  --task "sft:data_process" \
+  --use_gradient_checkpointing \
+  --dataset_num_workers 8 \
+  --find_unused_parameters
+
+accelerate launch --config_file examples/qwen_image/model_training/special/low_vram_training/deepspeed_zero3_cpuoffload.yaml examples/qwen_image/model_training/train.py \
+  --dataset_base_path "./models/train/Qwen-Image_lora-splited-cache" \
+  --max_pixels 1048576 \
+  --dataset_repeat 50 \
+  --model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors" \
+  --learning_rate 1e-4 \
+  --num_epochs 5 \
+  --remove_prefix_in_ckpt "pipe.dit." \
+  --output_path "./models/train/Qwen-Image_lora" \
+  --lora_base_model "dit" \
+  --lora_target_modules "to_q,to_k,to_v,add_q_proj,add_k_proj,add_v_proj,to_out.0,to_add_out,img_mlp.net.2,img_mod.1,txt_mlp.net.2,txt_mod.1" \
+  --lora_rank 32 \
+  --task "sft:train" \
+  --use_gradient_checkpointing \
+  --dataset_num_workers 8 \
+  --find_unused_parameters \
+  --initialize_model_on_cpu
+```
+
+其中，`accelerate` 和 `deepspeed` 的配置文件如下：
+
+```yaml
+compute_environment: LOCAL_MACHINE
+debug: true
+deepspeed_config:
+  deepspeed_config_file: examples/qwen_image/model_training/special/low_vram_training/ds_z3_cpuoffload.json
+  zero3_init_flag: true
+distributed_type: DEEPSPEED
+downcast_bf16: 'no'
+enable_cpu_affinity: false
+machine_rank: 0
+main_training_function: main
+num_machines: 1
+num_processes: 1
+rdzv_backend: static
+same_network: true
+tpu_env: []
+tpu_use_cluster: false
+tpu_use_sudo: false
+use_cpu: false
+```
+
+```json
+{
+    "fp16": {
+        "enabled": "auto",
+        "loss_scale": 0,
+        "loss_scale_window": 1000,
+        "initial_scale_power": 16,
+        "hysteresis": 2,
+        "min_loss_scale": 1
+    },
+    "bf16": {
+        "enabled": "auto"
+    },
+    "zero_optimization": {
+        "stage": 3,
+        "offload_optimizer": {
+            "device": "cpu",
+            "pin_memory": true
+        },
+        "offload_param": {
+            "device": "cpu",
+            "pin_memory": true
+        },
+        "overlap_comm": false,
+        "contiguous_gradients": true,
+        "sub_group_size": 1e9,
+        "reduce_bucket_size": 5e7,
+        "stage3_prefetch_bucket_size": 5e7,
+        "stage3_param_persistence_threshold": 1e5,
+        "stage3_max_live_parameters": 1e8,
+        "stage3_max_reuse_distance": 1e8,
+        "stage3_gather_16bit_weights_on_model_save": true
+    },
+    "activation_checkpointing": {
+        "partition_activations": false,
+        "cpu_checkpointing": false,
+        "contiguous_memory_optimization": false
+    },
+    "gradient_accumulation_steps": "auto",
+    "gradient_clipping": "auto",
+    "train_batch_size": "auto",
+    "train_micro_batch_size_per_gpu": "auto",
+    "wall_clock_breakdown": false
+}
+```

docs/zh/Training/Direct_Distill.md

@@ -77,7 +77,7 @@ distill_qwen/image.jpg,"精致肖像，水下少女，蓝裙飘逸，发丝轻
 这个样例数据集可以直接下载：
 
 ```shell
-modelscope download --dataset DiffSynth-Studio/example_image_dataset --local_dir ./data/example_image_dataset
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --local_dir ./data/diffsynth_example_dataset
 ```
 
 然后开始 LoRA 蒸馏加速训练：

docs/zh/index.rst

@@ -13,6 +13,7 @@
 
    Pipeline_Usage/Setup
    Pipeline_Usage/Model_Inference
+   Pipeline_Usage/Accelerated_Inference
    Pipeline_Usage/VRAM_management
    Pipeline_Usage/Model_Training
    Pipeline_Usage/Environment_Variables
@@ -29,6 +30,10 @@
    Model_Details/Z-Image
    Model_Details/Anima
    Model_Details/LTX-2
+   Model_Details/ERNIE-Image
+   Model_Details/JoyAI-Image
+   Model_Details/Stable-Diffusion
+   Model_Details/Stable-Diffusion-XL
 
 .. toctree::
    :maxdepth: 2

examples/anima/README.md

@@ -0,0 +1,3 @@
+English Document: https://diffsynth-studio-doc.readthedocs.io/en/latest/Model_Details/Anima.html
+
+中文文档：https://diffsynth-studio-doc.readthedocs.io/zh-cn/latest/Model_Details/Anima.html

examples/anima/model_training/full/anima-preview.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "anima/anima-preview/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch examples/anima/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata.csv \
+  --dataset_base_path data/diffsynth_example_dataset/anima/anima-preview \
+  --dataset_metadata_path data/diffsynth_example_dataset/anima/anima-preview/metadata.csv \
   --max_pixels 1048576 \
   --dataset_repeat 50 \
   --model_id_with_origin_paths "circlestone-labs/Anima:split_files/diffusion_models/anima-preview.safetensors,circlestone-labs/Anima:split_files/text_encoders/qwen_3_06b_base.safetensors,circlestone-labs/Anima:split_files/vae/qwen_image_vae.safetensors" \
@@ -11,4 +13,4 @@ accelerate launch examples/anima/model_training/train.py \
   --remove_prefix_in_ckpt "pipe.dit." \
   --output_path "./models/train/anima-preview_full" \
   --trainable_models "dit" \
-  --use_gradient_checkpointing
+  --use_gradient_checkpointing

examples/anima/model_training/lora/anima-preview.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "anima/anima-preview/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch examples/anima/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata.csv \
+  --dataset_base_path data/diffsynth_example_dataset/anima/anima-preview \
+  --dataset_metadata_path data/diffsynth_example_dataset/anima/anima-preview/metadata.csv \
   --max_pixels 1048576 \
   --dataset_repeat 50 \
   --model_id_with_origin_paths "circlestone-labs/Anima:split_files/diffusion_models/anima-preview.safetensors,circlestone-labs/Anima:split_files/text_encoders/qwen_3_06b_base.safetensors,circlestone-labs/Anima:split_files/vae/qwen_image_vae.safetensors" \
@@ -13,4 +15,4 @@ accelerate launch examples/anima/model_training/train.py \
   --lora_base_model "dit" \
   --lora_target_modules "" \
   --lora_rank 32 \
-  --use_gradient_checkpointing
+  --use_gradient_checkpointing

examples/ernie_image/model_inference/ERNIE-Image-Turbo.py

@@ -0,0 +1,25 @@
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image-Turbo", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=8,
+    cfg_scale=1.0,
+    sigma_shift=4.0,
+)
+image.save("output_turbo.jpg")

examples/ernie_image/model_inference/ERNIE-Image.py

@@ -0,0 +1,24 @@
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("output.jpg")

examples/ernie_image/model_inference_low_vram/ERNIE-Image-Turbo.py

@@ -0,0 +1,37 @@
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image-Turbo", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=8,
+    cfg_scale=1.0,
+    sigma_shift=4.0,
+)
+image.save("output_turbo.jpg")

examples/ernie_image/model_inference_low_vram/ERNIE-Image.py

@@ -0,0 +1,36 @@
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+import torch
+
+vram_config = {
+    "offload_dtype": torch.bfloat16,
+    "offload_device": "cpu",
+    "onload_dtype": torch.bfloat16,
+    "onload_device": "cpu",
+    "preparing_dtype": torch.bfloat16,
+    "preparing_device": "cuda",
+    "computation_dtype": torch.bfloat16,
+    "computation_device": "cuda",
+}
+
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device='cuda',
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors", **vram_config),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
+    ],
+    tokenizer_config=ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/"),
+    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
+)
+
+image = pipe(
+    prompt="一只黑白相间的中华田园犬",
+    negative_prompt="",
+    height=1024,
+    width=1024,
+    seed=42,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("output.jpg")

examples/ernie_image/model_training/full/ERNIE-Image.sh

@@ -0,0 +1,17 @@
+# Dataset: data/diffsynth_example_dataset/ernie_image/Ernie-Image-T2I/
+
+accelerate launch --config_file examples/ernie_image/model_training/full/accelerate_config_zero3.yaml \
+  examples/ernie_image/model_training/train.py \
+  --dataset_base_path data/diffsynth_example_dataset/ernie_image/Ernie-Image-T2I \
+  --dataset_metadata_path data/diffsynth_example_dataset/ernie_image/Ernie-Image-T2I/metadata.csv \
+  --max_pixels 1048576 \
+  --dataset_repeat 50 \
+  --model_id_with_origin_paths "PaddlePaddle/ERNIE-Image:transformer/diffusion_pytorch_model*.safetensors,PaddlePaddle/ERNIE-Image:text_encoder/model.safetensors,PaddlePaddle/ERNIE-Image:vae/diffusion_pytorch_model.safetensors" \
+  --learning_rate 1e-5 \
+  --num_epochs 2 \
+  --remove_prefix_in_ckpt "pipe.dit." \
+  --output_path "./models/train/Ernie-Image-T2I_full" \
+  --trainable_models "dit" \
+  --use_gradient_checkpointing \
+  --dataset_num_workers 8 \
+  --find_unused_parameters

examples/ernie_image/model_training/full/accelerate_config_zero3.yaml

@@ -0,0 +1,23 @@
+compute_environment: LOCAL_MACHINE
+debug: false
+deepspeed_config:
+  gradient_accumulation_steps: 1
+  offload_optimizer_device: none
+  offload_param_device: none
+  zero3_init_flag: true
+  zero3_save_16bit_model: true
+  zero_stage: 3
+distributed_type: DEEPSPEED
+downcast_bf16: 'no'
+enable_cpu_affinity: false
+machine_rank: 0
+main_training_function: main
+mixed_precision: bf16
+num_machines: 1
+num_processes: 8
+rdzv_backend: static
+same_network: true
+tpu_env: []
+tpu_use_cluster: false
+tpu_use_sudo: false
+use_cpu: false

examples/ernie_image/model_training/lora/ERNIE-Image.sh

@@ -0,0 +1,19 @@
+# Dataset: data/diffsynth_example_dataset/ernie_image/Ernie-Image-T2I/
+# Download: modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "ernie_image/Ernie-Image-T2I/*" --local_dir ./data/diffsynth_example_dataset
+
+accelerate launch examples/ernie_image/model_training/train.py \
+  --dataset_base_path data/diffsynth_example_dataset/ernie_image/Ernie-Image-T2I \
+  --dataset_metadata_path data/diffsynth_example_dataset/ernie_image/Ernie-Image-T2I/metadata.csv \
+  --max_pixels 1048576 \
+  --dataset_repeat 50 \
+  --model_id_with_origin_paths "PaddlePaddle/ERNIE-Image:transformer/diffusion_pytorch_model*.safetensors,PaddlePaddle/ERNIE-Image:text_encoder/model.safetensors,PaddlePaddle/ERNIE-Image:vae/diffusion_pytorch_model.safetensors" \
+  --learning_rate 1e-4 \
+  --num_epochs 5 \
+  --remove_prefix_in_ckpt "pipe.dit." \
+  --output_path "./models/train/Ernie-Image-T2I_lora" \
+  --lora_base_model "dit" \
+  --lora_target_modules "to_q,to_k,to_v,to_out.0" \
+  --lora_rank 32 \
+  --use_gradient_checkpointing \
+  --dataset_num_workers 8 \
+  --find_unused_parameters

examples/ernie_image/model_training/train.py

@@ -0,0 +1,129 @@
+import torch, os, argparse, accelerate
+from diffsynth.core import UnifiedDataset
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+from diffsynth.diffusion import *
+from diffsynth.core.data.operators import *
+os.environ["TOKENIZERS_PARALLELISM"] = "false"
+
+
+class ErnieImageTrainingModule(DiffusionTrainingModule):
+    def __init__(
+        self,
+        model_paths=None, model_id_with_origin_paths=None,
+        tokenizer_path=None,
+        trainable_models=None,
+        lora_base_model=None, lora_target_modules="", lora_rank=32, lora_checkpoint=None,
+        preset_lora_path=None, preset_lora_model=None,
+        use_gradient_checkpointing=True,
+        use_gradient_checkpointing_offload=False,
+        extra_inputs=None,
+        fp8_models=None,
+        offload_models=None,
+        device="cpu",
+        task="sft",
+    ):
+        super().__init__()
+        # Load models
+        model_configs = self.parse_model_configs(model_paths, model_id_with_origin_paths, fp8_models=fp8_models, offload_models=offload_models, device=device)
+        tokenizer_config = ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="tokenizer/") if tokenizer_path is None else ModelConfig(tokenizer_path)
+        self.pipe = ErnieImagePipeline.from_pretrained(torch_dtype=torch.bfloat16, device=device, model_configs=model_configs, tokenizer_config=tokenizer_config)
+        self.pipe = self.split_pipeline_units(task, self.pipe, trainable_models, lora_base_model)
+
+        # Training mode
+        self.switch_pipe_to_training_mode(
+            self.pipe, trainable_models,
+            lora_base_model, lora_target_modules, lora_rank, lora_checkpoint,
+            preset_lora_path, preset_lora_model,
+            task=task,
+        )
+
+        # Other configs
+        self.use_gradient_checkpointing = use_gradient_checkpointing
+        self.use_gradient_checkpointing_offload = use_gradient_checkpointing_offload
+        self.extra_inputs = extra_inputs.split(",") if extra_inputs is not None else []
+        self.task = task
+        self.task_to_loss = {
+            "sft:data_process": lambda pipe, inputs_shared, inputs_posi, inputs_nega: (inputs_shared, inputs_posi, inputs_nega),
+            "sft": lambda pipe, inputs_shared, inputs_posi, inputs_nega: FlowMatchSFTLoss(pipe, **inputs_shared, **inputs_posi),
+            "sft:train": lambda pipe, inputs_shared, inputs_posi, inputs_nega: FlowMatchSFTLoss(pipe, **inputs_shared, **inputs_posi),
+        }
+
+    def get_pipeline_inputs(self, data):
+        inputs_posi = {"prompt": data["prompt"]}
+        inputs_nega = {"negative_prompt": ""}
+        inputs_shared = {
+            "input_image": data["image"],
+            "height": data["image"].size[1],
+            "width": data["image"].size[0],
+            "cfg_scale": 1,
+            "rand_device": self.pipe.device,
+            "use_gradient_checkpointing": self.use_gradient_checkpointing,
+            "use_gradient_checkpointing_offload": self.use_gradient_checkpointing_offload,
+        }
+        inputs_shared = self.parse_extra_inputs(data, self.extra_inputs, inputs_shared)
+        return inputs_shared, inputs_posi, inputs_nega
+
+    def forward(self, data, inputs=None):
+        if inputs is None:
+            inputs = self.get_pipeline_inputs(data)
+        inputs = self.transfer_data_to_device(inputs, self.pipe.device, self.pipe.torch_dtype)
+        for unit in self.pipe.units:
+            inputs = self.pipe.unit_runner(unit, self.pipe, *inputs)
+        loss = self.task_to_loss[self.task](self.pipe, *inputs)
+        return loss
+
+
+def ernie_image_parser():
+    parser = argparse.ArgumentParser(description="ERNIE-Image training.")
+    parser = add_general_config(parser)
+    parser = add_image_size_config(parser)
+    parser.add_argument("--tokenizer_path", type=str, default=None, help="Path to tokenizer.")
+    return parser
+
+
+if __name__ == "__main__":
+    parser = ernie_image_parser()
+    args = parser.parse_args()
+    accelerator = accelerate.Accelerator(
+        gradient_accumulation_steps=args.gradient_accumulation_steps,
+        kwargs_handlers=[accelerate.DistributedDataParallelKwargs(find_unused_parameters=args.find_unused_parameters)],
+    )
+    dataset = UnifiedDataset(
+        base_path=args.dataset_base_path,
+        metadata_path=args.dataset_metadata_path,
+        repeat=args.dataset_repeat,
+        data_file_keys=args.data_file_keys.split(","),
+        main_data_operator=lambda x: x,
+        special_operator_map={
+            "image": ToAbsolutePath(args.dataset_base_path) >> LoadImage() >> ImageCropAndResize(args.height, args.width, args.max_pixels, 16, 16),
+        },
+    )
+    model = ErnieImageTrainingModule(
+        model_paths=args.model_paths,
+        model_id_with_origin_paths=args.model_id_with_origin_paths,
+        tokenizer_path=args.tokenizer_path,
+        trainable_models=args.trainable_models,
+        lora_base_model=args.lora_base_model,
+        lora_target_modules=args.lora_target_modules,
+        lora_rank=args.lora_rank,
+        lora_checkpoint=args.lora_checkpoint,
+        preset_lora_path=args.preset_lora_path,
+        preset_lora_model=args.preset_lora_model,
+        use_gradient_checkpointing=args.use_gradient_checkpointing,
+        use_gradient_checkpointing_offload=args.use_gradient_checkpointing_offload,
+        extra_inputs=args.extra_inputs,
+        fp8_models=args.fp8_models,
+        offload_models=args.offload_models,
+        task=args.task,
+        device=accelerator.device,
+    )
+    model_logger = ModelLogger(
+        args.output_path,
+        remove_prefix_in_ckpt=args.remove_prefix_in_ckpt,
+    )
+    launcher_map = {
+        "sft:data_process": launch_data_process_task,
+        "sft": launch_training_task,
+        "sft:train": launch_training_task,
+    }
+    launcher_map[args.task](accelerator, dataset, model, model_logger, args=args)

examples/ernie_image/model_training/validate_full/ERNIE-Image.py

@@ -0,0 +1,25 @@
+import torch
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+from diffsynth.core import load_state_dict
+
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
+    ],
+)
+
+state_dict = load_state_dict("./models/train/Ernie-Image-T2I_full/epoch-1.safetensors")
+pipe.dit.load_state_dict(state_dict)
+
+image = pipe(
+    prompt="a professional photo of a cute dog",
+    seed=0,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("image_full.jpg")
+print("Full validation image saved to image_full.jpg")

examples/ernie_image/model_training/validate_lora/ERNIE-Image.py

@@ -0,0 +1,25 @@
+import torch
+from diffsynth.pipelines.ernie_image import ErnieImagePipeline, ModelConfig
+from diffsynth.core.loader.file import load_state_dict
+
+pipe = ErnieImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="PaddlePaddle/ERNIE-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
+    ],
+)
+
+lora_state_dict = load_state_dict("./models/train/Ernie-Image-T2I_lora/epoch-4.safetensors", torch_dtype=torch.bfloat16, device="cuda")
+pipe.load_lora(pipe.dit, state_dict=lora_state_dict, alpha=1.0)
+
+image = pipe(
+    prompt="a professional photo of a cute dog",
+    seed=0,
+    num_inference_steps=50,
+    cfg_scale=4.0,
+)
+image.save("image_lora.jpg")
+print("LoRA validation image saved to image_lora.jpg")

examples/flux/README.md

@@ -0,0 +1,3 @@
+English Document: https://diffsynth-studio-doc.readthedocs.io/en/latest/Model_Details/FLUX.html
+
+中文文档：https://diffsynth-studio-doc.readthedocs.io/zh-cn/latest/Model_Details/FLUX.html

examples/flux/model_training/full/FLEX.2-preview.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLEX.2-preview/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch --config_file examples/flux/model_training/full/accelerate_config.yaml examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLEX.2-preview \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLEX.2-preview/metadata.csv \
   --max_pixels 1048576 \
   --dataset_repeat 200 \
   --model_id_with_origin_paths "ostris/Flex.2-preview:Flex.2-preview.safetensors,black-forest-labs/FLUX.1-dev:text_encoder/model.safetensors,black-forest-labs/FLUX.1-dev:text_encoder_2/*.safetensors,black-forest-labs/FLUX.1-dev:ae.safetensors" \

examples/flux/model_training/full/FLUX.1-Kontext-dev.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLUX.1-Kontext-dev/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch --config_file examples/flux/model_training/full/accelerate_config.yaml examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata_kontext.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLUX.1-Kontext-dev \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLUX.1-Kontext-dev/metadata.csv \
   --data_file_keys "image,kontext_images" \
   --max_pixels 1048576 \
   --dataset_repeat 400 \

examples/flux/model_training/full/FLUX.1-Krea-dev.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLUX.1-Krea-dev/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch --config_file examples/flux/model_training/full/accelerate_config.yaml examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLUX.1-Krea-dev \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLUX.1-Krea-dev/metadata.csv \
   --max_pixels 1048576 \
   --dataset_repeat 400 \
   --model_id_with_origin_paths "black-forest-labs/FLUX.1-Krea-dev:flux1-krea-dev.safetensors,black-forest-labs/FLUX.1-dev:text_encoder/model.safetensors,black-forest-labs/FLUX.1-dev:text_encoder_2/*.safetensors,black-forest-labs/FLUX.1-dev:ae.safetensors" \

examples/flux/model_training/full/FLUX.1-dev-AttriCtrl.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLUX.1-dev-AttriCtrl/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata_attrictrl.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLUX.1-dev-AttriCtrl \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLUX.1-dev-AttriCtrl/metadata.csv \
   --data_file_keys "image" \
   --max_pixels 1048576 \
   --dataset_repeat 100 \

examples/flux/model_training/full/FLUX.1-dev-Controlnet-Inpainting-Beta.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLUX.1-dev-Controlnet-Inpainting-Beta/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch --config_file examples/flux/model_training/full/accelerate_config.yaml examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata_controlnet_inpaint.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLUX.1-dev-Controlnet-Inpainting-Beta \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLUX.1-dev-Controlnet-Inpainting-Beta/metadata.csv \
   --data_file_keys "image,controlnet_image,controlnet_inpaint_mask" \
   --max_pixels 1048576 \
   --dataset_repeat 400 \

examples/flux/model_training/full/FLUX.1-dev-Controlnet-Union-alpha.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLUX.1-dev-Controlnet-Union-alpha/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch --config_file examples/flux/model_training/full/accelerate_config.yaml examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata_controlnet_canny.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLUX.1-dev-Controlnet-Union-alpha \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLUX.1-dev-Controlnet-Union-alpha/metadata.csv \
   --data_file_keys "image,controlnet_image" \
   --max_pixels 1048576 \
   --dataset_repeat 400 \

examples/flux/model_training/full/FLUX.1-dev-Controlnet-Upscaler.sh

@@ -1,6 +1,8 @@
+modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux/FLUX.1-dev-Controlnet-Upscaler/*" --local_dir ./data/diffsynth_example_dataset
+
 accelerate launch --config_file examples/flux/model_training/full/accelerate_config.yaml examples/flux/model_training/train.py \
-  --dataset_base_path data/example_image_dataset \
-  --dataset_metadata_path data/example_image_dataset/metadata_controlnet_upscale.csv \
+  --dataset_base_path data/diffsynth_example_dataset/flux/FLUX.1-dev-Controlnet-Upscaler \
+  --dataset_metadata_path data/diffsynth_example_dataset/flux/FLUX.1-dev-Controlnet-Upscaler/metadata.csv \
   --data_file_keys "image,controlnet_image" \
   --max_pixels 1048576 \
   --dataset_repeat 400 \