siliconflow/onediff

OneDiff: An out-of-the-box acceleration library for diffusion models.

66
/ 100
Established

Combines PyTorch graph compilation with custom GPU kernels to optimize diffusion model inference, supporting both Nexfort and OneFlow backends. Integrates seamlessly with Hugging Face Diffusers, ComfyUI, and Stable Diffusion WebUI through minimal code changes, while offering production-ready features like cached compilation and distributed inference to avoid recompilation overhead.

1,973 stars and 1,380 monthly downloads. Available on PyPI.

Maintenance 6 / 25
Adoption 17 / 25
Maturity 25 / 25
Community 18 / 25

How are scores calculated?

Stars

1,973

Forks

127

Language

Jupyter Notebook

License

Apache-2.0

Last pushed

Dec 04, 2025

Monthly downloads

1,380

Commits (30d)

0

Dependencies

4

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/siliconflow/onediff"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.