Commit History

Autor SHA1 Mensaxe Data
  royjhan a5f23d766e Merge branch 'main' into royh-batchembed hai 10 meses
  Roy Han 512e0a7bde Clean up hai 10 meses
  Roy Han aee25acb5b move normalization to go hai 10 meses
  Jeffrey Morgan 717f7229eb Do not shift context for sliding window models (#5368) hai 10 meses
  Roy Han 80c1a3f812 playing around with truncate stuff hai 10 meses
  Roy Han c111d8bb51 normalization hai 10 meses
  Roy Han 5213c12354 clean up hai 10 meses
  Roy Han 49e341147d add server function hai 10 meses
  Roy Han c406fa7a4c api/embed draft hai 10 meses
  Roy Han ff191d7cba Initial Draft hai 10 meses
  Michael Yang 9d91e5e587 remove confusing log message hai 10 meses
  Daniel Hiltgen fb9cdfa723 Fix server.cpp for the new cuda build macros hai 11 meses
  Jeffrey Morgan ead259d877 llm: fix seed value not being applied to requests (#4986) hai 10 meses
  Jeffrey Morgan 34f142797a llm: always add bos token to prompt (#4941) hai 11 meses
  Michael Yang 829ff87bd1 revert tokenize ffi (#4761) hai 11 meses
  Michael Yang de781b37c8 rm unused infill hai 11 meses
  Michael Yang 3e21799377 rm unused system prompt hai 11 meses
  Michael Yang 26a00a0410 use ffi for tokenizing/detokenizing hai 11 meses
  Michael Yang 714adb8bd1 bump (#4597) hai 11 meses
  Daniel Hiltgen b37b496a12 Wire up load progress hai 11 meses
  Sam e15307fdf4 feat: add support for flash_attn (#4120) hai 11 meses
  Michael Yang 58876091f7 log clean up hai 1 ano
  Daniel Hiltgen 920a4b0794 Merge remote-tracking branch 'upstream/main' into pr3702 hai 1 ano
  Michael Yang 44869c59d6 omit prompt and generate settings from final response hai 1 ano
  jmorganca fcf4d60eee llm: add back check for empty token cache hai 1 ano
  Jeffrey Morgan 18d9a7e1f1 update llama.cpp submodule to `f364eb6` (#4060) hai 1 ano
  Daniel Hiltgen 23d23409a0 Update llama.cpp (#4036) hai 1 ano
  ManniX-ITA c942e4a07b Fixed startup sequence to report model loading hai 1 ano
  Jeffrey Morgan 7c9792a6e0 Support unicode characters in model path (#3681) hai 1 ano
  Daniel Hiltgen 0a0e9f3e0f Apply 01-cache.diff hai 1 ano