On the compression of shallow non-causal ASR models using knowledge distillation and tied-and-reduced decoder for low-latency on-device speech recognition

Nagaraj Adiga, Jinhwan Park, Chintigari Shiva Kumar +4

Recently, the cascaded two-pass architecture has emerged as a strong contender for on-device automatic speech recognition (ASR). A cascade of causal and shallow non-causal encoders…