Cog Video https://github.com/THUDM/CogVideo

Installation step by step for local use

1.Install ComfyUI https://github.com/comfyanonymous/ComfyUI/releases/latest/download/ComfyUI_windows_portable_nvidia.7z

2. Open your directory with ComfyUi ----> ''custom nodes'' folder  and enter the command ''cmd.exe'' in the address bar of the window

In terminal git clone https://github.com/kijai/ComfyUI-CogVideoXWrapper 

and enter

3.Open ''Custom nodes/ComfyUI-CogVideoXWrapper'' folder, 

cmd.exe in address bar

pip install -r requirements.txt    and enter

4. Next download wokflow-file file https://openart.ai/workflows/seal_harmful_40/cogvideox-5b---golden-retriever/S9KV8o6Nbkw4irboTOyS

5.Install Custom Nodes Manager . Download this file to comfyui directory and double click

6.Open ComfUi and delete all nodes. Place your workflow-file on this empty space

7. Click Manager and Click Install missing custom nodes

8.Restart and ctrl+F5

9. Download text encoders https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main (t5xxlFp8......) and to .....ComfyUI_windows_portable\ComfyUI\models\clip folder

10. Select t5xxl_fp8_e4m3fn.safetensors

11.Select THUDM/CogVideoX-2b

12.Your prompt to video in Cog video text encode

12. Control auto generate - Randomise

And Queue Prompt

After will download the models first. If he does not download the model, then download it manually here, repeating the folder structure https://huggingface.co/THUDM/CogVideoX-2b/tree/main/transformer

model diffusion_pytorch_model.safetensors

This folder ....ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\models\CogVideo\CogVideo2B