mirror of
https://github.com/modelscope/DiffSynth-Studio.git
synced 2026-03-23 00:58:11 +00:00
update
This commit is contained in:
Artiprocher
20
examples/flux/model_training/validate_full/FLEX.2-preview.py
Normal file
20
examples/flux/model_training/validate_full/FLEX.2-preview.py
Normal file
@@ -0,0 +1,20 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
|
||||
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLEX.2-preview_full/epoch-0.safetensors")
|
||||
pipe.dit.load_state_dict(state_dict)
|
||||
|
||||
image = pipe(prompt="dog,white and brown dog, sitting on wall, under pink flowers", seed=0)
|
||||
image.save("image_FLEX.2-preview_full.jpg")
|
||||
@@ -0,0 +1,26 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-Kontext-dev_full/epoch-0.safetensors")
|
||||
pipe.dit.load_state_dict(state_dict)
|
||||
|
||||
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_full.jpg")
|
||||
@@ -0,0 +1,20 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
|
||||
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-Krea-dev_full/epoch-0.safetensors")
|
||||
pipe.dit.load_state_dict(state_dict)
|
||||
|
||||
image = pipe(prompt="a dog", seed=0)
|
||||
image.save("image_FLUX.1-Krea-dev_full.jpg")
|
||||
@@ -0,0 +1,21 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
|
||||
|
||||
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")
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-AttriCtrl_full/epoch-0.safetensors")
|
||||
pipe.value_controller.encoders[0].load_state_dict(state_dict)
|
||||
|
||||
image = pipe(prompt="a cat", seed=0, value_controller_inputs=0.1, rand_device="cuda")
|
||||
image.save("image_FLUX.1-dev-AttriCtrl_full.jpg")
|
||||
@@ -0,0 +1,31 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-Controlnet-Inpainting-Beta_full/epoch-0.safetensors")
|
||||
pipe.controlnet.models[0].load_state_dict(state_dict)
|
||||
|
||||
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_full.jpg")
|
||||
@@ -0,0 +1,31 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-Controlnet-Union-alpha_full/epoch-0.safetensors")
|
||||
pipe.controlnet.models[0].load_state_dict(state_dict)
|
||||
|
||||
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_full.jpg")
|
||||
@@ -0,0 +1,30 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-Controlnet-Upscaler_full/epoch-0.safetensors")
|
||||
pipe.controlnet.models[0].load_state_dict(state_dict)
|
||||
|
||||
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_full.jpg")
|
||||
@@ -0,0 +1,28 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-IP-Adapter_full/epoch-0.safetensors")
|
||||
pipe.ipadapter.load_state_dict(state_dict)
|
||||
|
||||
image = pipe(
|
||||
prompt="a dog",
|
||||
ipadapter_images=Image.open("data/example_image_dataset/1.jpg"),
|
||||
height=768, width=768,
|
||||
seed=0
|
||||
)
|
||||
image.save("image_FLUX.1-dev-IP-Adapter_full.jpg")
|
||||
@@ -0,0 +1,33 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig, ControlNetInput
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-InfiniteYou_full/epoch-0.safetensors")
|
||||
state_dict_projector = {i.replace("image_proj_model.", ""): state_dict[i] for i in state_dict if i.startswith("image_proj_model.")}
|
||||
pipe.image_proj_model.load_state_dict(state_dict_projector)
|
||||
state_dict_controlnet = {i.replace("controlnet.models.0.", ""): state_dict[i] for i in state_dict if i.startswith("controlnet.models.0.")}
|
||||
pipe.controlnet.models[0].load_state_dict(state_dict_controlnet)
|
||||
|
||||
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_full.jpg")
|
||||
@@ -0,0 +1,25 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
|
||||
|
||||
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/LoRA-Encoder-FLUX.1-Dev", origin_file_pattern="model.safetensors"),
|
||||
],
|
||||
)
|
||||
pipe.enable_lora_magic()
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev-LoRA-Encoder_full/epoch-0.safetensors")
|
||||
pipe.lora_encoder.load_state_dict(state_dict)
|
||||
|
||||
lora = ModelConfig(model_id="VoidOc/flux_animal_forest1", origin_file_pattern="20.safetensors")
|
||||
pipe.load_lora(pipe.dit, lora, hotload=True) # Use `pipe.clear_lora()` to drop the loaded LoRA.
|
||||
|
||||
image = pipe(prompt="", seed=0, lora_encoder_inputs=lora)
|
||||
image.save("image_FLUX.1-dev-LoRA-Encoder_full.jpg")
|
||||
20
examples/flux/model_training/validate_full/FLUX.1-dev.py
Normal file
20
examples/flux/model_training/validate_full/FLUX.1-dev.py
Normal file
@@ -0,0 +1,20 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
|
||||
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-dev_full/epoch-0.safetensors")
|
||||
pipe.dit.load_state_dict(state_dict)
|
||||
|
||||
image = pipe(prompt="a dog", seed=0)
|
||||
image.save("image_FLUX.1-dev_full.jpg")
|
||||
28
examples/flux/model_training/validate_full/Nexus-Gen.py
Normal file
28
examples/flux/model_training/validate_full/Nexus-Gen.py
Normal file
@@ -0,0 +1,28 @@
|
||||
import torch
|
||||
from PIL import Image
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/FLUX.1-NexusGen-Edit_full/epoch-0.safetensors")
|
||||
pipe.dit.load_state_dict(state_dict)
|
||||
|
||||
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=2.0, num_inference_steps=50,
|
||||
nexus_gen_reference_image=ref_image,
|
||||
height=512, width=512,
|
||||
)
|
||||
image.save("NexusGen-Edit_full.jpg")
|
||||
25
examples/flux/model_training/validate_full/Step1X-Edit.py
Normal file
25
examples/flux/model_training/validate_full/Step1X-Edit.py
Normal file
@@ -0,0 +1,25 @@
|
||||
import torch
|
||||
from diffsynth.pipelines.flux_image import FluxImagePipeline, ModelConfig
|
||||
from diffsynth import load_state_dict
|
||||
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"),
|
||||
],
|
||||
)
|
||||
state_dict = load_state_dict("models/train/Step1X-Edit_full/epoch-0.safetensors")
|
||||
pipe.dit.load_state_dict(state_dict)
|
||||
|
||||
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_full.jpg")
|
||||
Reference in New Issue
Block a user