Historique des commits

Auteur SHA1 Message Date
  Tri Dao 0705d2718d [Llama] Fix some tests, add tests for Llama 2 and CodeLlama il y a 1 an
  Tri Dao dfe29f5e2b [Gen] Don't use ft_attention, use flash_attn_with_kvcache instead il y a 1 an
  Tri Dao 913922cac5 [Gen] Refactor decoding function il y a 1 an
  Tri Dao 942fcbf046 [Rotary] Implement rotary in Triton il y a 1 an
  Tri Dao 0e8c46ae08 Run isort and black on test files il y a 1 an
  Tri Dao 4d87e4d875 Implement GPT-J il y a 1 an
  Tri Dao 78b7a1dc18 [OPT] Load fp16 weights on CPU before moving to GPU il y a 2 ans
  Tri Dao 88173a1aaf [FusedDense] Support relu, rename FusedDenseGeluDense -> FusedMLP il y a 2 ans