Histórico de Commits

Autor SHA1 Mensagem Data
  Michael Yang b732beba6a lint há 9 meses atrás
  Jeffrey Morgan c4cf8ad559 llm: avoid loading model if system memory is too small (#5637) há 9 meses atrás
  Jeffrey Morgan f8241bfba3 gpu: report system free memory instead of 0 (#5521) há 10 meses atrás
  Daniel Hiltgen 6f351bf586 review comments and coverage há 11 meses atrás
  Daniel Hiltgen fc37c192ae Refine CPU load behavior with system memory visibility há 11 meses atrás
  Daniel Hiltgen 30a7d7096c Bump VRAM buffer back up há 11 meses atrás
  Michael Yang 4736391bfb llm: add minimum based on layer size há 1 ano atrás
  Jeffrey Morgan f0c454ab57 gpu: add 512MiB to darwin minimum, metal doesn't have partial offloading overhead (#4068) há 1 ano atrás
  Daniel Hiltgen 34b9db5afc Request and model concurrency há 1 ano atrás
  Michael Yang 26df674785 scale graph based on gpu count há 1 ano atrás
  Michael Yang 41a272de9f darwin: no partial offloading if required memory greater than system há 1 ano atrás
  Michael Yang 7e33a017c0 partial offloading há 1 ano atrás
  Daniel Hiltgen be330174dd Allow setting max vram for workarounds há 1 ano atrás
  peanut256 a189810df6 Determine max VRAM on macOS using `recommendedMaxWorkingSetSize` (#2354) há 1 ano atrás
  Daniel Hiltgen 7427fa1387 Fix up the CPU fallback selection há 1 ano atrás
  Daniel Hiltgen 39928a42e8 Always dynamically load the llm server library há 1 ano atrás
  Daniel Hiltgen d88c527be3 Build multiple CPU variants and pick the best há 1 ano atrás
  Jeffrey Morgan c336693f07 calculate overhead based number of gpu devices (#1875) há 1 ano atrás
  Jeffrey Morgan 08f1e18965 Offload layers to GPU based on new model size estimates (#1850) há 1 ano atrás
  Jeffrey Morgan c7ea8f237e set `num_gpu` to 1 only by default on darwin arm64 (#1771) há 1 ano atrás
  Daniel Hiltgen a2ad952440 Fix windows system memory lookup há 1 ano atrás
  Daniel Hiltgen d966b730ac Switch windows build to fully dynamic há 1 ano atrás
  Daniel Hiltgen 7555ea44f8 Revamp the dynamic library shim há 1 ano atrás
  Daniel Hiltgen 6558f94ed0 Fix darwin intel build há 1 ano atrás
  Daniel Hiltgen 35934b2e05 Adapted rocm support to cgo based llama.cpp há 1 ano atrás