Commit History

Author SHA1 Message Date
  Jay Shah 50cb90aea6 comment out unimplemented kwargs from flash_attn_with_kvcache 4 months ago
  Jay Shah b5cac6d586 rebase with Is_local disabled temporarily 5 months ago
  Jay Shah 2472e5e0b4 add 'in principle' fp8 kv cache support 5 months ago
  Jay Shah e36e004cb3 change fp8 code path to allow for split kernel and kv cache without perf regression 5 months ago
  Jay Shah f53703b98d add split kv heuristic modifications 5 months ago
  Ganesh Bikshandi 099ca28488 add gqa decoding logic. 5 months ago
  Jay Shah a06f1f970b add to kv cache api 6 months ago
  Jay Shah 535b8279bb complete gqa parallel changes for non-causal 6 months ago
  Ganesh Bikshandi 84e31c22d6 adding python interface. 6 months ago
  Ying Zhang dff976a84a fixes 6 months ago
  Ying Zhang 7b4e68e04f hopper local attention 6 months ago
  Ying Zhang db80387343 Add seqused_q in fwd / bwd and seqused_k in bwd. 6 months ago
  Charlene Yang bdf733be55 Add q, k, v descales to FA3 interface (#1210) 5 months ago
  jayhshah c92ca63268 FA3 FP8 qkv descales + restore max offset for h128 causal + added sync for producer WG (#1173) 6 months ago
  Tri Dao bafe253042 [FA3] Bwd 7 months ago
  Tri Dao 3aae9c18c1 Revert "Changes For FP8 (#1075)" 7 months ago
  ganeshcolfax 1899c970c8 Changes For FP8 (#1075) 7 months ago
  Ying Zhang dfe1a59e4b Add var-seq-len to FA3 fp16 / bf16 fwd (#1072) 7 months ago
  youkaichao ef3e358a25 remove lambda (#1056) 7 months ago
  Tri Dao 7f67966cc7 FA3 initial code release 7 months ago