História revízii

Autor SHA1 Správa Dátum
  Josh Yan c800a67f1b cast 10 mesiacov pred
  Josh Yan dfc62648f3 cast 10 mesiacov pred
  Josh Yan 24e8292e94 new changes 10 mesiacov pred
  Josh Yan c63b4ecbf7 quantize 10 mesiacov pred
  Daniel Hiltgen b51e3b63ac Statically link c++ and thread lib 10 mesiacov pred
  jmorganca a08f20d910 release: remove unwanted mingw dll.a files 10 mesiacov pred
  jmorganca 6cea036027 Revert "llm: only statically link libstdc++" 10 mesiacov pred
  jmorganca 5796bfc401 llm: only statically link libstdc++ 10 mesiacov pred
  jmorganca f1a379aa56 llm: statically link pthread and stdc++ dependencies in windows build 10 mesiacov pred
  Jeffrey Morgan 5304b765b2 llm: put back old include dir (#5507) 10 mesiacov pred
  Jeffrey Morgan 78fb33dd07 fix typo in cgo directives in `llm.go` (#5501) 10 mesiacov pred
  Jeffrey Morgan 8f8e736b13 update llama.cpp submodule to `d7fd29f` (#5475) 10 mesiacov pred
  Michael Yang 829ff87bd1 revert tokenize ffi (#4761) 11 mesiacov pred
  Jeffrey Morgan 763bb65dbb use `int32_t` for call to tokenize (#4738) 11 mesiacov pred
  Michael Yang bf54c845e9 vocab only 11 mesiacov pred
  Michael Yang 26a00a0410 use ffi for tokenizing/detokenizing 1 rok pred
  Michael Yang 01811c176a comments 1 rok pred
  Michael Yang 9685c34509 quantize any fp16/fp32 model 1 rok pred
  Hernan Martinez 86e67fc4a9 Add import declaration for windows,arm64 to llm.go 1 rok pred
  Michael Yang 9502e5661f cgo quantize 1 rok pred
  Daniel Hiltgen 58d95cc9bd Switch back to subprocessing for llama.cpp 1 rok pred
  Michael Yang 91b3e4d282 update memory calcualtions 1 rok pred
  Michael Yang d338d70492 refactor model parsing 1 rok pred
  Patrick Devine 1b272d5bcd change `github.com/jmorganca/ollama` to `github.com/ollama/ollama` (#3347) 1 rok pred
  Jeffrey Morgan f9cd55c70b disable gpu for certain model architectures and fix divide-by-zero on memory estimation 1 rok pred
  Daniel Hiltgen 6c5ccb11f9 Revamp ROCm support 1 rok pred
  Daniel Hiltgen a1dfab43b9 Ensure the libraries are present 1 rok pred
  Jeffrey Morgan 4458efb73a Load all layers on `arm64` macOS if model is small enough (#2149) 1 rok pred
  Daniel Hiltgen fedd705aea Mechanical switch from log to slog 1 rok pred
  Michael Yang eaed6f8c45 add max context length check 1 rok pred