Login / Signup

TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-Device ASR Models.

Yuan ShangguanHaichuan YangDanni LiChunyang WuYassir FathullahDilin WangAyushi DalmiaRaghuraman KrishnamoorthiOzlem KalinliJunteng JiaJay MahadeokarXin LeiMike SeltzerVikas Chandra
Published in: ICASSP (2024)
Keyphrases
  • cost effective
  • low cost
  • image compression
  • nearest neighbor
  • information retrieval
  • case study
  • artificial neural networks
  • lightweight
  • video coding
  • complex systems
  • statistical model
  • recurrent neural networks