nunchaku-tech

nunchaku-tech/nunchaku

[ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models

3.0K stars 168 forks 3.0K watchers Python Apache License 2.0

comfyui diffusion-models flux genai iclr iclr2025 lora mlsys quantization

View on GitHub Website

2 Open Issues Need Help Last updated: Sep 14, 2025

Open Issues Need Help

View All on GitHub

[Bug] Fix tensor shape mismatch in fused_qkv_norm_rottary for non-square resolutions 3 months ago

good first issue

nunchaku-tech/nunchaku

3.0K

[ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models

Python

#comfyui#diffusion-models#flux#genai#iclr#iclr2025#lora#mlsys#quantization

Flux.1 Krea nunchaku 是否支持原始Flux.1的lora模型？ 3 months ago

good first issue

nunchaku-tech/nunchaku

3.0K

[ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models

Python

#comfyui#diffusion-models#flux#genai#iclr#iclr2025#lora#mlsys#quantization