1 paper
Vyom Agarwal, Mokshda Gangrade, Siddharth Pal +1
Large automatic speech recognition (ASR) models such as Whisper must be deployed across hardware with widely varying memory and inference-speed constraints. We present a compressio…