comfyui installation guide and animaldiff usage

Cloud disk files

If the network cannot be downloaded, I also found the corresponding network disk file
link: https://pan.quark.cn/s/19c83810e915

comfyui installation

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

img

https://github.com/comfyanonymous/ComfyUI/releases/download/latest/ComfyUI_windows_portable_nvidia_cu121_or_cpu.7z

Unzip

After unzipping, go to the following page, cmd, git clone and download the following two plug-ins first (search my previous article if git is not installed)

img

Plugin Manager

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

ComfyUI Chinese plug-in

https://github.com/AIGODLIKE/AIGODLIKE-ComfyUI-Translation.git

img

Click run_nvidia_gpu.bat,

After downloading the dependent files, click Settings in the upper right corner and select Chinese

img

Shared model

Maybe the sd used before, now copying the model to comfyui may waste space. The following method can solve the problem.

Copy extra_model_paths.yaml.example to the extra_model_paths.yaml file and open it in Notepad

Change base_path to the sd root directory

Modify controlnet to extend the controlnet model directory and save

img

img

How to load a workflow

Let me teach you an idea. It doesn’t matter if you don’t know how to build a workflow. The official website of comfyui provides a lot of workflows, such as the picture and example below. Generally, JSON workflow files or pictures are given. You can download them directly and drag them into comfyui automatically. The same workflow is generated; (picture drag-in, workflow load)

There are other plug-in workflows. You can go directly to the official code and there are similar files, such as animaldiff. Download the workflow and load it directly.

https://github.com/comfyanonymous/ComfyUI

img

https://github.com/ArtVentureX/comfyui-animatediff

img

animatediff example

When loading the animatediff workflow, you may be prompted that nodes are missing and need to be installed. Just restart after installation. You can also download the v2 model and motion model. See my last article and put it in the corresponding position below.

img

img

Then it’s time to adjust some parameters. I believe it shouldn’t be difficult for those who have played with SD.

Guess you like

Origin blog.csdn.net/weixin_62403633/article/details/133994238