https://github.com/comfyanonymous/ComfyUI/releases
ComfyUI_windows_portable_nvidia.7zD:\ComfyUI)进入解压目录,双击:
run_nvidia_gpu.bat首次启动会:
Running on http://127.0.0.1:8188浏览器访问:
http://127.0.0.1:8188git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUIpython -m venv venv
venv\Scripts\activatepip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
pip install -r requirements.txtpython main.py浏览器访问:
http://127.0.0.1:8188ComfyUI 本身不含模型,需要手动放模型。
ComfyUI/
└─ models/
├─ checkpoints/ # 大模型(SD1.5 / SDXL)
├─ vae/
├─ loras/
├─ controlnet/示例:
v1-5-pruned-emaonly.ckptsd_xl_base_1.0.safetensors放入:
ComfyUI\models\checkpointsrun_nvidia_gpu.bat--lowvram 或 --medvrampython main.py --lowvramrun_cpu.bat如果你愿意,我可以:
你现在的显卡和系统是啥?