Skip to content

ComfyUI-MotionDiff

ComfyUI-MotionDiff integrates several motion and pose estimation technologies such as 4DHuman, MotionGPT, MDM, MotionDiffuse, and ReMoDiffuse into ComfyUI, enhancing its capabilities in generating and rendering 3D poses from various inputs. It offers functionalities for estimating 3D poses, generating motion frames, and converting motion data to SMPL format for visualization, with support for various parameters to adjust the output quality and details. The project also includes features for saving and loading SMPL data, allowing for easy export to 3DCGI software like Blender, Unreal Engine, and Unity.

Tags

Animation * CLIP * CLIPConditioning * CameraControl * Conditioning * Face * MotionData * SMPL * SMPLModel

Repo info

  • Repo url: https://github.com/Fannovel16/ComfyUI-MotionDiff.git
  • Commit hash: f270ac206ba4e26c06144349252935df84004ab0

Licenses

  • MIT: LICENSE
  • BSD-3-Clause, LicenseRef-scancode-proprietary-license: ORIGINAL_LICENSE
  • LicenseRef-scancode-unknown-license-reference: clean_ch.py
  • Apache-2.0: custom_mmpkg/custom_mmseg/models/losses/dice_loss.py
  • MIT: custom_mmpkg/custom_mmseg/models/losses/lovasz_loss.py
  • MIT: motiondiff_modules/mGPT/data/humanml/common/quaternion.py
  • CC-BY-NC-SA-4.0: motiondiff_modules/spectre/LICENSE
  • MIT: motiondiff_modules/spectre/external/ibug/face_detection/retina_face/py_cpu_nms.py
  • Apache-2.0: motiondiff_modules/spectre/tracker/face_tracker.py

Visit licenses page for the details