Nodes Browser

ComfyDeploy: How ComfyUI-BiRefNet-lite works in ComfyUI?

What is ComfyUI-BiRefNet-lite?

This repository packages the latest BiRefNet model as a ComfyUI node for use, supporting chunked loading on both CPU and GPU, as well as model caching features.

How to install it in ComfyDeploy?

Head over to the machine page

  1. Click on the "Create a new machine" button
  2. Select the Edit build steps
  3. Add a new step -> Custom Node
  4. Search for ComfyUI-BiRefNet-lite and select it
  5. Close the build step dialig and then click on the "Save" button to rebuild the machine
<h1 align="center">ComfyUI-BiRefNet-lite</h1> <p align="center"> <br> <font size=5>中文 | <a href="README_EN.md">English</a></font> </p>

介绍

本仓库将BiRefNet最新模型封装为ComfyUI节点来使用,并支持cpu和gpu分块加载,模型缓存特性。<br>

特性

特性1: 支持cpu和gpu分块加载<br> 在开启cuda的情况下,指定cpu_size,可将模型部分加载到cpu上 slot<br> 特性2: 模型缓存<br> 特性3: 多种裁剪方式<br> 支持putalpha, naive, alpha_matting三种裁剪方式 slot<br> <br>

News

  • Nov 19, 2023: 增加mask_precision_threshold参数,控制mask的精度阈值,默认为0.1<br>

安装

方法1:

  1. 进入节点目录, ComfyUI/custom_nodes/
  2. git clone https://github.com/rubi-du/ComfyUI-BiRefNet-lite.git
  3. cd ComfyUI-BiRefNet-lite
  4. pip install -r requirements.txt
  5. 重启ComfyUI

方法2:

直接下载节点源码包,然后解压到custom_nodes目录下,最后重启ComfyUI

方法3:

通过ComfyUI-Manager安装,搜索“ComfyUI-BiRefNet-lite”进行安装

使用

示例工作流放置在ComfyUI-BiRefNet-lite/workflow<br/>

加载模型有两种方式可选:一是自动下载并加载远程模型,二是加载本地模型(此时需将load_local_model设为true,并将local_model_path设置为指向models/birefnet目录下的本地模型路径,例如BiRefNet文件夹)<br/>


模型下载地址:<br/> BiRefNet: https://huggingface.co/ZhengPeng7/BiRefNet/tree/main<br/> BiRefNet_lite-2K: https://huggingface.co/ZhengPeng7/BiRefNet_lite-2K/tree/main<br/> BiRefNet-portrait: https://huggingface.co/ZhengPeng7/BiRefNet-portrait/tree/main<br/> BiRefNet-matting: https://huggingface.co/ZhengPeng7/BiRefNet-matting/tree/main<br/> RMBG-2.0: https://huggingface.co/briaai/RMBG-2.0/tree/main<br/>


工作流workflow.json的使用<br/>

plot


工作流video_workflow.json的使用<br/> workflow地址

感谢

感谢BiRefNet仓库的所有作者 ZhengPeng7/BiRefNet

部分代码参考了 MoonHugo/ComfyUI-BiRefNet-Hugo 感谢!