PackQViT: Faster Sub-8-bit Vision Transformers via Full and Packed Quantization on the Mobile

Devs

PackQViT: Faster Sub-8-bit Vision Transformers via Full and Packed Quantization on the Mobile | Read Paper on Bytez