comfyui

# 部署

github 主页：https://github.com/comfyanonymous/ComfyUI

windowns 便捷版：https://github.com/comfyanonymous/ComfyUI/releases

## 执行更新

打开update目录

执行update_comfyui_and_python_dependencies.bat

## 修改启动脚本

.\python_embeded\python.exe -s ComfyUI\main.py  --listen 0.0.0.0 --port 8781 --windows-standalone-build --cuda-device 2

pause

## 部署管理器

https://github.com/ltdrdata/ComfyUI-Manager

# 安装xformers

python.exe -m pip install -U xformers --index-url https://download.pytorch.org/whl/cu124

# 安装apex

git clone https://github.com/NVIDIA/apex

cd apex

python -m pip install -v --disable-pip-version-check --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./apex --no-build-isolation

python -m pip install pytorch-extension

# 手动安装节点依赖

.\python_embeded\python.exe -m pip install -r ComfyUI\custom_nodes\ComfyUI-Sana\requirements.txt

# 节点精选

## pulid 人脸一致性生成

https://github.com/balazik/ComfyUI-PuLID-Flux

python.exe -m pip install insightface-0.7.3-cp311-cp311-win_amd64.whl

python_embeded\python.exe -m pip install -r filterpy\requirements.txt

python.exe -m pip install onnxruntime

python.exe -m pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

python.exe -m pip install ftfy

python.exe -m pip install timm

# ComfyUI_CatVTON_Wrapper换装 依赖（ComfyUI_LayerStyle）

https://github.com/chflame163/ComfyUI_CatVTON_Wrapper

# 蒙版 ComfyUI_LayerStyle

https://github.com/chflame163/ComfyUI_LayerStyle

如遇到异常：onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be load

解决方法：https://blog.csdn.net/IT_ZB/article/details/140442328

python.exe -m pip uninstall onnxruntime-gpu

python.exe -m pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

..\..\..\python_embeded\python.exe -s -m pip install -r requirements.txt

.\repair_dependency.bat

python download_models.py --all

.\python_embeded\python.exe ComfyUI\custom_nodes\comfyui-faceless-node\download_models.py --all

VAE Encode: img + vae  → latent

VAE Decode: latent + vae  → img

Upscale Latent: latent → latent

Empty Latent Image: new Latent

CLIP Text Encode：txt + clip model → txt tensor

Sampler ：txt tensor + latent + sd model → latent

LoraLoader: sd model + lora model → sd model

CLIP Vision Encode + unCLIPConditioning: img + clip version model + txt tensor → txt tensor

CLIP Text Encode (Prompt): condition → condition(置零)

temp、imput、output内容多会占用上传带宽

标识的序号-面板中的序号

0>2 （lora训练，）

1>3(sdwebui 8860 8861, comfyui 8783生产)

2>4(confyui 8789 开发生产环境8789)

3>5(photomaker 7860，不启用)（comfyui 8782 生产）

4>0(iclight 8960)

5>1(iclight 8961，不启用)（comfyui 8781 测试）

# 插件下载不了

1.重启comfyui

2.重启vpn

3.设置git 全局忽略证书

git config http.sslVerify"false"

git config --global http.sslVerify false