arXiv - CSCL: "Sparse Progressive Distillation: Resolving Overfi…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

Sparse Progressive Distillation: Resolving Overfitting under Pretrain-and-Finetune Paradigm. (arXiv:2110.08190v4 [cs.CL] UPDATED)

http://arxiv.org/abs/2110.08190 #arXiv #NLProc

Jan 18, 2023, 03:07 · · arxiv-cscl · · ·

Sign in to participate in the conversation