README.md

---
license: other
license_name: tencent-hunyuan-community
license_link: https://huggingface.co/tencent/HunyuanVideo/blob/6204ad6aea1a77ff5aba337c88278bb9500eb37/LICENSE
base_model: hunyuanvideo-community/HunyuanVideo
pipeline_tag: text-to-video
tags:
- hunyuanvideo
- diffusionart
- arxiv:2610.12572
---

# DiffusionART · HunyuanVideo

[Paper](https://arxiv.org/abs/2610.12572) ·
[Code](https://github.com/vanzll/DiffusionART) ·
[Project Page](https://vanzll.github.io/diffusionart-project/) ·
[Demo](https://vanzll.github.io/diffusionart-project/demo.html)

HunyuanVideo-13B fine-tuned with **DiffusionART** for text-to-video generation.

This repository contains the **full transformer weights**, not a LoRA or a
complete pipeline. The inference entry below downloads the base model's VAE
and text encoders automatically.

## Inference

Use the included [infer.py](https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/blob/main/infer.py).
Expand the setup once, then run the example from the same directory.

<details>
<summary>Environment setup</summary>

Linux, Python 3.12, a CUDA 12.1-compatible driver and toolkit (`nvcc`), and
approximately 100 GiB of available GPU memory and disk space for the verified
default path. The installer uses a pinned public DanceGRPO inference runtime.

```bash
export RUN_DIR="${RUN_DIR:-$PWD/diffusionart-t2v-demo}"
export HF_HOME="$RUN_DIR/cache/hf"
export PIP_CACHE_DIR="$RUN_DIR/pip-cache"
export TMPDIR="$RUN_DIR/tmp"
export XDG_CACHE_HOME="$RUN_DIR/xdg-cache"
mkdir -p "$RUN_DIR" "$HF_HOME" "$PIP_CACHE_DIR" "$TMPDIR" "$XDG_CACHE_HOME"
cd "$RUN_DIR"
python3.12 -m venv bootstrap
bootstrap/bin/python -m pip install 'huggingface-hub==0.26.1'
bootstrap/bin/huggingface-cli login

bootstrap/bin/python - <<'PY'
from huggingface_hub import snapshot_download
snapshot_download(
    "vanzl/DiffusionART-HunyuanVideo",
    revision="b9f83d8fc677be078cf20e75e2149a5f69059220",
    allow_patterns=["infer.py", "install.sh", "requirements-inference.txt"],
    local_dir="release",
)
PY
bash release/install.sh "$RUN_DIR/venv" "$RUN_DIR/runtime"
```

The fixed revision preserves the released weights and inference entry.
Hugging Face login is used by that entry even though this model is public.

</details>

```bash
venv/bin/python - <<'PY'
import runpy
import sys
import torch

# Use the attention backend from the successful inference checks.
torch.backends.cuda.enable_cudnn_sdp(False)
sys.argv = [
    "infer.py",
    "--revision", "b9f83d8fc677be078cf20e75e2149a5f69059220",
    "--prompt", "A white horse gallops through shallow water along a sunlit beach.",
    "--output", "diffusionart.mp4",
    "--cache-dir", "cache",
]
runpy.run_path("release/infer.py", run_name="__main__")
PY
```

Defaults: **640×640 · 53 frames · 8 FPS · 30 sampling steps**.
The script uses the verified shift, guidance, precision and VAE settings.
Change `--prompt` and `--output` for another video; an existing output is never overwritten.

## Examples

Selected outputs from this model. See the [Demo](https://vanzll.github.io/diffusionart-project/demo.html)
for the full gallery and pretrained comparisons.

<table>
<tr>
<td><video controls muted loop playsinline preload="none" width="280" poster="https://vanzll.github.io/diffusionart-project/assets/hf-video-00113-after.jpg" src="https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/resolve/main/examples/00113-diffusionart.mp4"></video><p>A white horse gallops through shallow water along a sunlit beach.</p></td>
<td><video controls muted loop playsinline preload="none" width="280" poster="https://vanzll.github.io/diffusionart-project/assets/hf-video-00139-after.jpg" src="https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/resolve/main/examples/00139-diffusionart.mp4"></video><p>The camera approaches a glass pavilion surrounded by autumn trees.</p></td>
</tr>
<tr>
<td><video controls muted loop playsinline preload="none" width="280" poster="https://vanzll.github.io/diffusionart-project/assets/hf-video-00106-after.jpg" src="https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/resolve/main/examples/00106-diffusionart.mp4"></video><p>The camera follows a winding river through a valley of autumn trees.</p></td>
<td><video controls muted loop playsinline preload="none" width="280" poster="https://vanzll.github.io/diffusionart-project/assets/hf-video-00148-after.jpg" src="https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/resolve/main/examples/00148-diffusionart.mp4"></video><p>Sunlight sparkles through a rotating crystal prism on a white surface.</p></td>
</tr>
</table>

## License

Subject to the [Tencent Hunyuan Community License](https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/blob/main/LICENSE.txt)
and [Notice](https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/blob/main/Notice).
Training provenance and weight hashes are recorded in [release.json](https://huggingface.co/vanzl/DiffusionART-HunyuanVideo/blob/main/release.json).

## Citation

```bibtex
@misc{wan2026diffusionartefficientstablescalable,
  title={DiffusionART: Efficient, Stable and Scalable Reinforcement of Diffusion Models},
  author={Zhenglin Wan and Jiayu Chen and Zhengfeiyang Zhang and Pengfei Zhou and Wangbo Zhao and Fangyuan Kong and Pengfei Wan and Kun Gai and Yang You},
  year={2026},
  eprint={2610.12572},
  archivePrefix={arXiv},
  primaryClass={cs.LG},
  url={https://arxiv.org/abs/2610.12572},
}
```