Description#
MotionLab (or simply MoLab) is a toolset designed to facilitate the generation and automatic in-betweening of human animation. It was developed as part of the Kollani research project, which focused on improving workflows of animation studios by integrating AI technologies.
It features:
- Text To Motion: Describe the motion you want to see
- In-Betweening: Input your keyposes and let AI fill the gaps
- Motion Composition: Compose sequences and generate transitions
Technologies#
- PyTorch: Used in the inference worker for running the motion generation models.
- CondMDI: The base motion generation project by Cohan et al. that MoLab builds upon.
- FastAPI: Used for the backend API to handle inference requests.
- Godot: The main MoLab interface and interaction logic is built using Godot, providing both a desktop and web interface.
- Maya+PyQt: A Plugin allows inference from within the Maya environment, enabling artists to work directly with their existing tools.
