准备工作
1.安装ComfyUI
2.安装python 3.10+
3.安装modelscope:pip install modelscope
进入comfyUI安装工作流自定义node
cd ComfyUI/custom_nodes
git clone https://github.com/Lightricks/ComfyUI-LTXVideo
git clone https://github.com/kosinkadink/ComfyUI-VideoHelperSuite
下载模型
1.下载LTX模型安装到ComfyUI/models/checkpoint:
modelscope download --model AI-ModelScope/LTX-Video ltx-video-2b-v0.9.safetensors --local_dir .
2.下载encoder安装到:ComfyUI/models/text_encoder
modelscope download --model Comfy-Org/stable-diffusion-3.5-fp8 text_encoder/t5xxl_fp16.safetensors --local_dir .
modelscope download --model CloseGPT/PixArt-XL-2-1024-MS(下载整个目录)
加载工作流
工作流地址:https://github.com/Lightricks/ComfyUI-LTXVideo?tab=readme-ov-file
文生视频:
图生视频: