david/flash-attention

réplica de https://github.com/Dao-AILab/flash-attention

Autor	SHA1 Mensaxe	Data
Tri Dao	a157cc8c9b [FT] Implement MQA/GQA	hai 1 ano
Tri Dao	62e9814466 [Rotary] Make sure frequency calculation is in fp32	hai 1 ano
Tri Dao	48bc6eacd6 [Gen] Add rotary base as an argument to FT attention kernel	hai 1 ano
Tri Dao	a01d1213d7 [Gen] Add kernel from FasterTransformer for benchmarking	%!s(int64=2) %!d(string=hai) anos