shipfeedAI news, curated daily

23:54:43 CET
20 MAY23:54:43shipfeed⋯
pull to refreshlast sync
Just in — 30 new
§ research · storyline

Full finetuning with pretraining optimizer reduces forgetting

Full finetuning with pretraining optimizer reduces forgetting

May 7 · · primary fetch1 sourceupdated May 7 ·

This storyline groups 1 article from 1 source. The originating feed didn’t ship an excerpt — open any link below to read the piece.

read full article on arxiv.org ↗
§ sources1 publication · timeline below
  1. arxiv.orgOptimizer-Model Consistency: Full Finetuning with the Same Optimizer as Pretraining Forgets Lessprimary