SATA

KO | EN

gitlite — search

#python #android #deep-learning #linux #docker #macos #java #nodejs #cli #kubernetes #awesome-list #security #openai #open-source

(★ 11)

[ICML 2026] Official repository for Semantic-Aware Motion Encoding for Topology-Agnostic Character Animation.

SATA Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# SATA: Semantic-Aware Motion Encoding for Topology-Agnostic Character Animation

[![Project Page](https://img.shields.io/badge/Project-Page-blue)](https://zzysteve.github.io/project_pages/sata.html)
[![arXiv](https://img.shields.io/badge/arXiv-b31b1b)](https://arxiv.org/abs/2605.27055)
[![Code](https://img.shields.io/badge/Code-GitHub-black)](https://github.com/zzysteve/SATA)
[![Hugging Face Demo](https://img.shields.io/badge/Hugging%20Face-Demo-ffcc4d?logo=huggingface)](https://huggingface.co/spaces/SteveZh/SATA)

Official repository for **Semantic-Aware Motion Encoding for Topology-Agnostic Character Animation**.

SATA is a semantic-aware, topology-agnostic motion representation framework for heterogeneous character animation. It learns a unified latent motion manifold across diverse skeletal topologies and supports motion reconstruction, text-to-motion generation, and zero-shot cross-species retargeting.

We release the core implementation and demo code. We recommend trying the
hosted demo first on [Hugging Face](https://huggingface.co/spaces/SteveZh/SATA).

We are actively working on releasing the remaining code.

## Quick Start

Follow these steps to run inference without downloading and processing the training data.

### Clone Repository

Clone this repository with submodules so the external code is checked out at the
paths expected by the demo and artifact scripts:

```shell
git clone --recurse-submodules https://github.com/zzysteve/SATA.git
cd SATA
```

If you already cloned the repository without submodules, initialize them from the
repository root:

```shell
git submodule sync --recursive
git submodule update --init --recursive
```

The submodule directory names are intentionally fixed by `.gitmodules`:

```text
src/fairmotion
src/mdm
src/momask-preenc
```

These paths do not follow the remote GitHub repository names. Do not rename them
locally; the code and artifact paths expect the directories above.

### Install

```shell
conda create -n sata python=3.11
conda activate sata
bash install.sh
```

### Artifacts

Download all required artifacts from the repository root:

```shell
python scripts/download_artifacts.py --all
```

To download only model checkpoints or only demo assets:

```shell
python scripts/download_artifacts.py --models
python scripts/download_artifacts.py --assets
```

Model checkpoints are distributed through Hugging Face:

```text
https://huggingface.co/SteveZh/sata_models
```

Demo BVH/assets are distributed as
[`sata_demo_assets.zip`](https://drive.google.com/file/d/1nHqYT9zZfvbXsXAIpfR3YZYWPk8YvKEt/view?usp=sharing)
through Google Drive.

If Google Drive blocks automated downloads or the downloaded file is not a
valid zip, download `sata_demo_assets.zip` manually from the link above, place it
in the repository root, then extract it and verify the artifacts:

```shell
unzip sata_demo_assets.zip
python scripts/download_artifacts.py --check
```

## Run Demo

```shell
python app.py
```

## Links

- [Project page](https://zzysteve.github.io/project_pages/sata.html)
- [Paper](https://arxiv.org/abs/2605.27055)
- [Hugging Face demo](https://huggingface.co/spaces/SteveZh/SATA)
- [Model checkpoints](https://huggingface.co/SteveZh/sata_models)

## TODO

- [x] Release SATA encoder-decoder inference code.
- [x] Add demos for encoding/decoding and retargeting.
- [ ] Add the data processing pipeline.
- [ ] Release SATA encoder-decoder training code.
- [ ] Add a text-to-motion demo with diverse skeletons.

## Citation

```
@inproceedings{zhang2026sata,
  title     = {Semantic-Aware Motion Encoding for Topology-Agnostic Character Animation},
  author    = {Zongye Zhang and Yuzhuo Cui and Qingjie Liu and Yunhong Wang},
  booktitle = {Proceedings of the International Conference on Machine Learning},
  year      = {2026},
  note      = {To appear. arXiv:2605.27055},
  url       = {https://arxiv.org/abs/2605.27055},
}
```

## Contact

For questions about the project, please contact the authors or open an issue.

## License

The original SATA code in this repository is released under the Apache License
2.0. See [LICENSE](LICENSE) and [NOTICE](NOTICE).

The following directories are Git submodules and remain governed by their
respective upstream licenses:

```text
src/fairmotion       https://github.com/zzysteve/SATA-fairmotion.git
src/mdm              https://github.com/zzysteve/SATA-motion-diffusion-model.git
src/momask-preenc    https://github.com/zzysteve/SATA-momask-codes.git
```

Please refer to the license files in each submodule or upstream repository
before using or redistributing those components. Files or assets that carry
their own copyright or license notices remain governed by those notices. Model
checkpoints, datasets, demo assets, body models, and other downloaded artifacts
may have separate terms and are not covered by the top-level Apache License
unless explicitly stated by their providers.

SATA

SATA Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

SATA

SATA Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README