Histórico de Commits

Autor SHA1 Mensagem Data
  Tri Dao 6e8b25e426 Refactor há 2 meses atrás
  Ying Zhang 1c9717d699 address comments há 2 meses atrás
  Ying Zhang dff976a84a fixes há 3 meses atrás
  Ying Zhang 7b4e68e04f hopper local attention há 3 meses atrás
  Ying Zhang db80387343 Add seqused_q in fwd / bwd and seqused_k in bwd. há 3 meses atrás
  jayhshah c92ca63268 FA3 FP8 qkv descales + restore max offset for h128 causal + added sync for producer WG (#1173) há 3 meses atrás
  Tri Dao bafe253042 [FA3] Bwd há 4 meses atrás
  jayhshah 5018ac6ac5 Fp8 kernel with "in-kernel" transpose of V in producer (#1100) há 4 meses atrás
  Tri Dao 3aae9c18c1 Revert "Changes For FP8 (#1075)" há 4 meses atrás
  ganeshcolfax 1899c970c8 Changes For FP8 (#1075) há 4 meses atrás
  Ying Zhang dfe1a59e4b Add var-seq-len to FA3 fp16 / bf16 fwd (#1072) há 4 meses atrás
  Cameron Shinn cb516f855b Remove torchlib dependency from cpp files (#1083) há 4 meses atrás
  Tri Dao 74b0761ff7 [FA3] BF16 forward há 5 meses atrás
  Tri Dao 7f67966cc7 FA3 initial code release há 5 meses atrás