david/flash-attention

spogulis no https://github.com/Dao-AILab/flash-attention

Autors	SHA1 Ziņojums	Datums
Tri Dao	71befc19e1 [Loss] Use flash_attn.losses.cross_entropy.CrossEntropyLoss	1 gadu atpakaļ
Tri Dao	dff68c2b22 Add smoothing for CrossEntropyParallel, rename to CrossEntropyLoss	2 gadi atpakaļ
Tri Dao	0bf5e50038 Release training code	2 gadi atpakaļ