EmpathicRobotics/FineVideo-VLA-Agent

## Work In Progress
- We are creating a CC-BY-SA permissive world model dataset

- Human 3D-pose b-spline action tokens dataset inspired by https://arxiv.org/abs/2506.06072

- Based on the awesome [FineVideo](https://huggingface.co/datasets/HuggingFaceFV/finevideo) dataset. See https://huggingface.co/blog/fine-video

- We aim to cleanup and extract as many human pose, including facial points, hands and feet as possible, and

- Train a model to retarget to various simulated and physical platforms, incluiding the H1 unitree.

### Sample Using Tokens -> Actions Generation
<video controls autoplay loop muted width="100%">
<source src="https://huggingface.co/datasets/EmpathicRobotics/FineVideo-VLA-Agent/resolve/main/sample.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>



If you want to help, please reach out with a message in the [community](https://huggingface.co/datasets/EmpathicRobotics/FineVideo-VLA-Agent/discussions) tab above.