Scaling Laws for Sparsely-Connected Foundation Models
— Aran Komatsuzaki (@arankomatsuzaki) September 18, 2023
Identifies the first scaling law describing the relationship between weight sparsity, number of non zero parameters, and amount of training datahttps://t.co/z7EmvXXgus pic.twitter.com/actodjm1by
No comments:
Post a Comment