Motius MotionCLR HumanML3D is a packaged checkpoint for the MotionCLR model, designed for generating human motion data from text prompts using the HumanML3D dataset. The repository provides a self-contained artifact that includes the denoising network, a frozen OpenAI CLIP ViT-B/32 text encoder, HumanML3D statistics, source revision, and license attribution. The model is intended for use with the Motius MotionCLR pipeline, enabling users to perform text-to-motion inference.
The tool outputs unnormalized HumanML3D motion sequences at 20 frames per second, with each output represented as a (T, 263) tensor. Users can generate motions by providing textual descriptions, such as 'a person walks forward and waves,' along with the desired sequence length. The repository includes instructions for loading the model and running inference using the provided pipeline.
0. The official checkpoint and related assets are sourced from the IDEA-Research/MotionCLR project. 5 billion parameters.
This tool is positioned within the class of text-to-motion generation models, specifically leveraging the HumanML3D dataset and the CLIP text encoder for its functionality.
In the Other AI space, Motius Motionclr Humanml3d takes a focused approach. It enables developers to generate human motion sequences from text descriptions using pretrained models. It is built as an open-source project for AI researchers and developers working on motion generation. Motius Motionclr Humanml3d is open source under the Open Source license. Motius Motionclr Humanml3d is available on the web and the command line.
Behind Motius Motionclr Humanml3d is ZeyuLing, and the product first shipped in 2026. Development happens publicly on GitHub with 68 commits in the last 90 days. Key capabilities include text-to-motion, motion generation, and humanML3D support.
Latest indexed changes and source events
ZeyuLing/motius-motionclr-humanml3d verified by the PulseGate indexer
