FlashAttention 2: making Transformers 800% faster w/o approximation - with Tri Dao of Together AI
1 tracks • 5406 plays • 0 favorites
FlashAttention 2: making Transformers 800% faster w/o approximation - with Tri Dao of Together AI
TheLazyMakoti
0:0054:35
Success & Inspiration
Episodes
1
FlashAttention 2: making Transformers 800% faster w/o approximation - with Tri Dao of Together AI
TheLazyMakoti