DiffSynth-Studio/Qwen-Image-Blockwise-ControlNet-Canny

1---2license: apache-2.03---4# Qwen-Image Image Structure Control Model5 6![](./assets/title.png)7 8## Model Introduction9 10This model is an image structure control model trained based on [Qwen-Image](https://www.modelscope.cn/models/Qwen/Qwen-Image), with the ControlNet architecture. It can control the structure of generated images using edge detection (Canny) maps. The training framework is built on [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio), and the dataset used is [BLIP3o](https://modelscope.cn/datasets/BLIP3o/BLIP3o-60k).11 12## Result Demonstration13 14|Canny Edge Map|Generated Image 1|Generated Image 2|15|-|-|-|16|![](./assets/canny_3.png)|![](./assets/image_3_1.png)|![](./assets/image_3_2.png)|17|![](./assets/canny_2.png)|![](./assets/image_2_1.png)|![](./assets/image_2_2.png)|18|![](./assets/canny_1.png)|![](./assets/image_1_1.png)|![](./assets/image_1_2.png)|19 20## Inference Code21```22git clone https://github.com/modelscope/DiffSynth-Studio.git  23cd DiffSynth-Studio24pip install -e .25```26 27```python28from diffsynth.pipelines.qwen_image import QwenImagePipeline, ModelConfig, ControlNetInput29from PIL import Image30import torch31from modelscope import dataset_snapshot_download32 33 34pipe = QwenImagePipeline.from_pretrained(35    torch_dtype=torch.bfloat16,36    device="cuda",37    model_configs=[38        ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors"),39        ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="text_encoder/model*.safetensors"),40        ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),41        ModelConfig(model_id="DiffSynth-Studio/Qwen-Image-Blockwise-ControlNet-Canny", origin_file_pattern="model.safetensors"),42    ],43    tokenizer_config=ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="tokenizer/"),44)45 46dataset_snapshot_download(47    dataset_id="DiffSynth-Studio/example_image_dataset",48    local_dir="./data/example_image_dataset",49    allow_file_pattern="canny/image_1.jpg"50)51controlnet_image = Image.open("data/example_image_dataset/canny/image_1.jpg").resize((1328, 1328))52```53 54prompt = "A little dog with shiny, soft fur and lively eyes, set in a spring courtyard with cherry blossoms falling, creating a beautiful and warm atmosphere."55image = pipe(56    prompt, seed=0,57    blockwise_controlnet_inputs=[ControlNetInput(image=controlnet_image)]58)59image.save("image.jpg")60```