/ViDiT-Q

[ICLR'25] ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video Generation

Primary LanguagePython

Watchers