Historique des commits

Auteur SHA1 Message Date
  AlpinDale 696f2cd59c add phi3_small support with blocksparse attention il y a 6 mois
  AlpinDale 0c15965621 fix fp8 kv il y a 6 mois
  AlpinDale a94de94c44 refactor: combine the prefill and decode into a single API (#553) il y a 6 mois
  AlpinDale 50b7c13db0 refactor: attention selector (#552) il y a 6 mois
  AlpinDale d11d68f4e6 switch to vllm-flash-attn il y a 6 mois
  AlpinDale 8b56dc4347 dict -> torch.Tensor for blocks_to_swap il y a 6 mois
  AlpinDale 3a0d1c7705 add get_name method to attention backends il y a 6 mois
  AlpinDale 21ce19b3ea blocks_to_copy dict -> torch.Tensor il y a 6 mois
  AlpinDale 2351a0e2cd feat: FlashInfer backend for decoding phase (#548) il y a 6 mois