david/flash-attention

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

Auteur	SHA1 Message	Date
Tri Dao	94657af3e8 Add option for not doing intra-WG overlapping of gemm and softmax	il y a 1 mois
Tri Dao	fe412d6b36 Redo rotary when contiguous	il y a 1 mois
Tri Dao	b2d3fe92ff Move rotary to a separate file	il y a 1 mois
Tri Dao	4d00645c76 Implement appending new KV to KV cache	il y a 1 mois
Tri Dao	d00b88ee05 Move PagedKV to a separate file	il y a 1 mois