Commit History

Auteur SHA1 Bericht Datum
  AlpinDale 8d77c69cbd feat: support image processor and add llava example 7 maanden geleden
  AlpinDale e4ea3da1ad fix: tensor parallel with embedding model 7 maanden geleden
  AlpinDale de62ceb18c refactor: eliminate parallel worker per-step task scheduling overhead 8 maanden geleden
  AlpinDale a94de94c44 refactor: combine the prefill and decode into a single API (#553) 8 maanden geleden
  AlpinDale 50b7c13db0 refactor: attention selector (#552) 8 maanden geleden
  AlpinDale be8154a8a0 feat: proper embeddings API with e5-mistral-7b support 8 maanden geleden