update

2026-03-22 16:50:47 +00:00 · 2025-11-19 20:22:21 +08:00
parent 6ad8d73717
commit eeb55a0ce6
88 changed files with 3113 additions and 78 deletions
--- a/examples/flux/model_training/validate_lora/FLEX.2-preview.py
+++ b/examples/flux/model_training/validate_lora/FLEX.2-preview.py
@@ -0,0 +1,18 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="ostris/Flex.2-preview", origin_file_pattern="Flex.2-preview.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLEX.2-preview_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(prompt="dog,white and brown dog, sitting on wall, under pink flowers", seed=0)
+image.save("image_FLEX.2-preview_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-Kontext-dev.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-Kontext-dev.py
@@ -0,0 +1,24 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-Kontext-dev", origin_file_pattern="flux1-kontext-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-Kontext-dev_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="Make the dog turn its head around.",
+    kontext_images=Image.open("data/example_image_dataset/2.jpg").resize((768, 768)),
+    height=768, width=768,
+    seed=0
+)
+image.save("image_FLUX.1-Kontext-dev_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-Krea-dev.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-Krea-dev.py
@@ -0,0 +1,18 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-Krea-dev", origin_file_pattern="flux1-krea-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-Krea-dev_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(prompt="a dog", seed=0)
+image.save("image_FLUX.1-Krea-dev_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-AttriCtrl.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-AttriCtrl.py
@@ -0,0 +1,19 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+        ModelConfig(model_id="DiffSynth-Studio/AttriCtrl-FLUX.1-Dev", origin_file_pattern="models/brightness.safetensors")
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-AttriCtrl_lora/epoch-3.safetensors", alpha=1)
+
+image = pipe(prompt="a cat", seed=0, value_controller_inputs=0.1, rand_device="cuda")
+image.save("image_FLUX.1-dev-AttriCtrl_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-Controlnet-Inpainting-Beta.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-Controlnet-Inpainting-Beta.py
@@ -0,0 +1,29 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+        ModelConfig(model_id="alimama-creative/FLUX.1-dev-Controlnet-Inpainting-Beta", origin_file_pattern="diffusion_pytorch_model.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-Controlnet-Inpainting-Beta_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="a cat sitting on a chair, wearing sunglasses",
+    controlnet_inputs=[ControlNetInput(
+        image=Image.open("data/example_image_dataset/inpaint/image_1.jpg"),
+        inpaint_mask=Image.open("data/example_image_dataset/inpaint/mask.jpg"),
+        scale=0.9
+    )],
+    height=1024, width=1024,
+    seed=0, rand_device="cuda",
+)
+image.save("image_FLUX.1-dev-Controlnet-Inpainting-Beta_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-Controlnet-Union-alpha.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-Controlnet-Union-alpha.py
@@ -0,0 +1,29 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+        ModelConfig(model_id="InstantX/FLUX.1-dev-Controlnet-Union-alpha", origin_file_pattern="diffusion_pytorch_model.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-Controlnet-Union-alpha_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="a dog",
+    controlnet_inputs=[ControlNetInput(
+        image=Image.open("data/example_image_dataset/canny/image_1.jpg"),
+        scale=0.9,
+        processor_id="canny",
+    )],
+    height=768, width=768,
+    seed=0, rand_device="cuda",
+)
+image.save("image_FLUX.1-dev-Controlnet-Union-alpha_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-Controlnet-Upscaler.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-Controlnet-Upscaler.py
@@ -0,0 +1,28 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+        ModelConfig(model_id="jasperai/Flux.1-dev-Controlnet-Upscaler", origin_file_pattern="diffusion_pytorch_model.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-Controlnet-Upscaler_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="a dog",
+    controlnet_inputs=[ControlNetInput(
+        image=Image.open("data/example_image_dataset/upscale/image_1.jpg"),
+        scale=0.9
+    )],
+    height=768, width=768,
+    seed=0, rand_device="cuda",
+)
+image.save("image_FLUX.1-dev-Controlnet-Upscaler_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-EliGen.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-EliGen.py
@@ -0,0 +1,33 @@
+import torch
+from PIL import Image
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+    ],
+)
+
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-EliGen_lora/epoch-4.safetensors", alpha=1)
+
+entity_prompts = ["A beautiful girl", "sign 'Entity Control'", "shorts", "shirt"]
+global_prompt = "A beautiful girl wearing shirt and shorts in the street,  holding a sign 'Entity Control'"
+masks = [Image.open(f"data/example_image_dataset/eligen/{i}.png").convert('RGB') for i in range(len(entity_prompts))]
+# generate image
+image = pipe(
+    prompt=global_prompt,
+    cfg_scale=1.0,
+    num_inference_steps=50,
+    embedded_guidance=3.5,
+    seed=42,
+    height=1024,
+    width=1024,
+    eligen_entity_prompts=entity_prompts,
+    eligen_entity_masks=masks,
+)
+image.save(f"EliGen_lora.png")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-IP-Adapter.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-IP-Adapter.py
@@ -0,0 +1,26 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+        ModelConfig(model_id="InstantX/FLUX.1-dev-IP-Adapter", origin_file_pattern="ip-adapter.bin"),
+        ModelConfig(model_id="google/siglip-so400m-patch14-384"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-IP-Adapter_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="dog,white and brown dog, sitting on wall, under pink flowers",
+    ipadapter_images=Image.open("data/example_image_dataset/1.jpg"),
+    height=768, width=768,
+    seed=0
+)
+image.save("image_FLUX.1-dev-IP-Adapter_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev-InfiniteYou.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev-InfiniteYou.py
@@ -0,0 +1,28 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+        ModelConfig(model_id="ByteDance/InfiniteYou", origin_file_pattern="infu_flux_v1.0/aes_stage2/image_proj_model.bin"),
+        ModelConfig(model_id="ByteDance/InfiniteYou", origin_file_pattern="infu_flux_v1.0/aes_stage2/InfuseNetModel/*.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev-InfiniteYou_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="a man with a red hat",
+    controlnet_inputs=[ControlNetInput(
+        image=Image.open("data/example_image_dataset/infiniteyou/image_1.jpg"),
+    )],
+    height=1024, width=1024,
+    seed=0, rand_device="cuda",
+)
+image.save("image_FLUX.1-dev-InfiniteYou_lora.jpg")
--- a/examples/flux/model_training/validate_lora/FLUX.1-dev.py
+++ b/examples/flux/model_training/validate_lora/FLUX.1-dev.py
@@ -0,0 +1,18 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="flux1-dev.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-dev_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(prompt="a dog", seed=0)
+image.save("image_FLUX.1-dev_lora.jpg")
--- a/examples/flux/model_training/validate_lora/Nexus-Gen.py
+++ b/examples/flux/model_training/validate_lora/Nexus-Gen.py
@@ -0,0 +1,26 @@
+import torch
+from PIL import Image
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="DiffSynth-Studio/Nexus-GenV2", origin_file_pattern="model*.safetensors"),
+        ModelConfig(model_id="DiffSynth-Studio/Nexus-GenV2", origin_file_pattern="edit_decoder.bin"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder/model.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="text_encoder_2/*.safetensors"),
+        ModelConfig(model_id="black-forest-labs/FLUX.1-dev", origin_file_pattern="ae.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/FLUX.1-NexusGen-Edit_lora/epoch-4.safetensors", alpha=1)
+
+ref_image = Image.open("data/example_image_dataset/nexus_gen/image_1.png").convert("RGB")
+prompt = "Add a pair of sunglasses."
+image = pipe(
+    prompt=prompt, negative_prompt="",
+    seed=42, cfg_scale=1.0, num_inference_steps=50,
+    nexus_gen_reference_image=ref_image,
+    height=512, width=512,
+)
+image.save("NexusGen-Edit_lora.jpg")
--- a/examples/flux/model_training/validate_lora/Step1X-Edit.py
+++ b/examples/flux/model_training/validate_lora/Step1X-Edit.py
@@ -0,0 +1,23 @@
+import torch
+from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
+from PIL import Image
+
+
+pipe = FluxImagePipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="Qwen/Qwen2.5-VL-7B-Instruct"),
+        ModelConfig(model_id="stepfun-ai/Step1X-Edit", origin_file_pattern="step1x-edit-i1258.safetensors"),
+        ModelConfig(model_id="stepfun-ai/Step1X-Edit", origin_file_pattern="vae.safetensors"),
+    ],
+)
+pipe.load_lora(pipe.dit, "models/train/Step1X-Edit_lora/epoch-4.safetensors", alpha=1)
+
+image = pipe(
+    prompt="Make the dog turn its head around.",
+    step1x_reference_image=Image.open("data/example_image_dataset/2.jpg").resize((768, 768)),
+    height=768, width=768, cfg_scale=6,
+    seed=0
+)
+image.save("image_Step1X-Edit_lora.jpg")