TY  - RPRT
TI  - Efficiently Train ASR Models that Memorize Less and Perform Better with Per-core Clipping
AU  - Lun Wang
AU  - Om Thakkar
AU  - Zhong Meng
AU  - Nicole Rafidi
AU  - Rohit Prabhavalkar
AU  - Arun Narayanan
PY  - 2024
UR  - https://arxiv.org/abs/2406.02004
ID  - 2406.02004
ER  -