DynamicPose-ComfyUI

DynamicPose-ComfyUI
★ 5

姿态提取视频合成UNet加载姿势引导
DynamicPose-ComfyUI 提供基于姿态的动态视频/图像合成管线节点集合，包含姿态提取、参考/去噪UNet加载与姿势引导采样，便于将姿态序列生成连续帧。
💡 将姿态序列转换为连续的视频帧
🍴 1 Forks💻 Python🔄 2024-09-11
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/e98a62d17551
example/example.png
📄 README

ComfyUI-Dynamicpose

Install

  • Clone this repo into the Your ComfyUI root directory\ComfyUI\custom_nodes\ and install dependent Python packages:

    • “`bash

    cd Your_ComfyUI_root_directory\ComfyUI\custom_nodes\

    git clone https://github.com/Bin-sam/DynamicPose-ComfyUI.git

    pip install -r requirements.txt

    “`

  • Download pre-trained models:
  • stable-diffusion-v1-5_unet
  • Moore-AnimateAnyone Pre-trained Models
  • DWpose model download links are under title “DWPose for ControlNet”.
  • Above models need to be put under folder pretrained_weights as follow:

    • “`text

    ./pretrained_weights/

    |– denoising_unet.pth

    |– motion_module.pth

    |– pose_guider.pth

    |– reference_unet.pth

    |– rtmpose

    | |– rtmw-x_simcc-cocktail14_pt-ucoco_270e-384×288-f840f204_20231122.pth

    | |– rtmw-x_8xb320-270e_cocktail14-384×288.py

    | |– rtmdet_m_640-8xb32_coco-person.py

    | |– rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth

    | |– rtmpose-x_8xb256-700e_coco-384×288.py

    | |– rtmpose-x_simcc-body7_pt-body7_700e-384×288-71d7b7e9_20230629.pth

    |– DWPose

    | |– dw-ll_ucoco_384.onnx

    | |– yolox_l.onnx

    “`

  • Download clip image encoder (e.g. sd-image-variations-diffusers ) and put it under Your_ComfyUI_root_directory\ComfyUI\models\clip_vision
  • Download vae (e.g. sd-vae-ft-mse) and put it under Your_ComfyUI_root_directory\ComfyUI\models\vae
  • Download stable-diffusion-v1-5 (e.g. stable-diffusion-v1-5) and put it under Your_ComfyUI_root_directory\ComfyUI\models\diffusers

    • Usage

  • Using pipeline examples

    • this workflow is Pose_to_video