# 3D动画、人物肖像 本模型使用 [**Qwen-Image-i2L**](https://modelscope.ai/models/DiffSynth-Studio/Qwen-Image-i2L) 从 2 张图像生成。 ## 安装 ```bash git clone https://github.com/modelscope/DiffSynth-Studio.git cd DiffSynth-Studio pip install -e . ``` ## 使用本模型 ```python from modelscope import snapshot_download from diffsynth.pipelines.qwen_image import QwenImagePipeline, ModelConfig import torch vram_config = { "offload_dtype": "disk", "offload_device": "disk", "onload_dtype": torch.bfloat16, "onload_device": "cpu", "preparing_dtype": torch.bfloat16, "preparing_device": "cuda", "computation_dtype": torch.bfloat16, "computation_device": "cuda", } pipe = QwenImagePipeline.from_pretrained( torch_dtype=torch.bfloat16, device="cuda", model_configs=[ ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="transformer/diffusion_pytorch_model*.safetensors", **vram_config), ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="text_encoder/model*.safetensors", **vram_config), ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config), ], tokenizer_config=ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="tokenizer/"), vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5, ) lora_path = snapshot_download(model_id="diffsynth-i2L-gallery/Qwen-Image-3Danimation-portrait-5a2b") pipe.load_lora(pipe.dit, f"{lora_path}/model.safetensors") image = pipe("一位非洲裔女性3D动画角色，面带微笑，穿着职业装或传统印花服饰，背景为纯色简洁影棚风格", seed=0, height=1024, width=1024, num_inference_steps=50) image.save("image.jpg") ``` ## 相关链接 - i2L 模型：https://modelscope.ai/models/DiffSynth-Studio/Qwen-Image-i2L - 代码仓库：https://github.com/modelscope/DiffSynth-Studio ## License Apache-2.0