sayakpaul/diffusers-torchao

End-to-end recipes for optimizing diffusion models with torchao and diffusers (inference and FP8 training).

46
/ 100
Emerging

Combines `torch.compile()` with `torchao`'s quantization schemes (INT8, FP8) to accelerate diffusion model inference, achieving 53.88% speedup on Flux.1-Dev and 27.33% on CogVideoX-5b. Provides end-to-end recipes for both inference optimization and experimental FP8 training, with serialization strategies to reduce framework overhead. Integrates directly into the Hugging Face `diffusers` pipeline as an official quantization backend, supporting automatic quantization (`autoquant`) with minimal code changes.

397 stars.

No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 10 / 25

How are scores calculated?

Stars

397

Forks

16

Language

Python

License

Apache-2.0

Last pushed

Jan 08, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/sayakpaul/diffusers-torchao"

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