theluyuan/DiffSynth-Studio
1
0
Fork 0
mirror of https://github.com/modelscope/DiffSynth-Studio.git synced 2026-03-24 10:18:12 +00:00
Code Issues Packages Projects Releases Wiki Activity

qwen_image layercontrol v2

Browse Source
This commit is contained in:
mi804
2026-02-24 15:19:16 +08:00
parent 288bbc7128
commit ee73a29885
10 changed files with 171 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
examples/qwen_image/model_inference/Qwen-Image-Layered-Control-V2.py Normal file
View File

@@ -0,0 +1,43 @@
from diffsynth.pipelines.qwen_image import QwenImagePipeline, ModelConfig
from modelscope import dataset_snapshot_download
from PIL import Image
import torch
pipe = QwenImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="Qwen/Qwen-Image-Layered", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),
ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="text_encoder/model*.safetensors"),
ModelConfig(model_id="Qwen/Qwen-Image-Layered", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, ModelConfig(model_id="DiffSynth-Studio/Qwen-Image-Layered-Control-V2", origin_file_pattern="model.safetensors"))
dataset_snapshot_download(
dataset_id="DiffSynth-Studio/example_image_dataset",
local_dir="./data/example_image_dataset",
allow_file_pattern="layer_v2/*.png"
)
prompt = "Text 'APRIL'"
input_image = Image.open("data/example_image_dataset/layer_v2/image_1.png").convert("RGBA").resize((1024, 1024))
image = pipe(
prompt, seed=0,
height=1024, width=1024,
layer_input_image=input_image, layer_num=0,
num_inference_steps=10, cfg_scale=4,
)
image[0].save("image_prompt.png")
mask_image = Image.open("data/example_image_dataset/layer_v2/mask_2.png").convert("RGBA").resize((1024, 1024))
input_image = Image.open("data/example_image_dataset/layer_v2/image_2.png").convert("RGBA").resize((1024, 1024))
image = pipe(
prompt, seed=0,
height=1024, width=1024,
layer_input_image=input_image, layer_num=0,
context_image=mask_image,
num_inference_steps=10, cfg_scale=1.0,
)
image[0].save("image_mask.png")