arXiv - CSCL: "Revisiting Intermediate Layer Distillation for Co…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

Revisiting Intermediate Layer Distillation for Compressing Language Models: An Overfitting Perspective. (arXiv:2302.01530v1 [cs.CL])

http://arxiv.org/abs/2302.01530 #arXiv #NLProc

Feb 06, 2023, 03:06 · · arxiv-cscl · · ·

Sign in to participate in the conversation