Bringing Nunchaku 4-bit Diffusion Inference to Diffusers
中文摘要
Nunchaku 将 4 位扩散推理引入 Diffusers 库，在保证生成质量的前提下，显著降低了显存占用并大幅提升了推理速度。
English Summary
Nunchaku brings efficient 4-bit diffusion inference to Diffusers, significantly reducing memory usage and boosting generation speed while maintaining high-quality outputs for generative AI models.