تاریخچه Commit ها

نویسنده SHA1 پیام تاریخ
  Daniel Hiltgen d470ebe78b Add Jetson cuda variants for arm 11 ماه پیش
  Michael Yang b732beba6a lint 9 ماه پیش
  Jeffrey Morgan c4cf8ad559 llm: avoid loading model if system memory is too small (#5637) 9 ماه پیش
  Jeffrey Morgan f8241bfba3 gpu: report system free memory instead of 0 (#5521) 10 ماه پیش
  Daniel Hiltgen 6f351bf586 review comments and coverage 11 ماه پیش
  Daniel Hiltgen fc37c192ae Refine CPU load behavior with system memory visibility 11 ماه پیش
  Daniel Hiltgen 30a7d7096c Bump VRAM buffer back up 11 ماه پیش
  Michael Yang 4736391bfb llm: add minimum based on layer size 1 سال پیش
  Jeffrey Morgan f0c454ab57 gpu: add 512MiB to darwin minimum, metal doesn't have partial offloading overhead (#4068) 1 سال پیش
  Daniel Hiltgen 34b9db5afc Request and model concurrency 1 سال پیش
  Michael Yang 26df674785 scale graph based on gpu count 1 سال پیش
  Michael Yang 41a272de9f darwin: no partial offloading if required memory greater than system 1 سال پیش
  Michael Yang 7e33a017c0 partial offloading 1 سال پیش
  Daniel Hiltgen be330174dd Allow setting max vram for workarounds 1 سال پیش
  peanut256 a189810df6 Determine max VRAM on macOS using `recommendedMaxWorkingSetSize` (#2354) 1 سال پیش
  Daniel Hiltgen 7427fa1387 Fix up the CPU fallback selection 1 سال پیش
  Daniel Hiltgen 39928a42e8 Always dynamically load the llm server library 1 سال پیش
  Daniel Hiltgen d88c527be3 Build multiple CPU variants and pick the best 1 سال پیش
  Jeffrey Morgan c336693f07 calculate overhead based number of gpu devices (#1875) 1 سال پیش
  Jeffrey Morgan 08f1e18965 Offload layers to GPU based on new model size estimates (#1850) 1 سال پیش
  Jeffrey Morgan c7ea8f237e set `num_gpu` to 1 only by default on darwin arm64 (#1771) 1 سال پیش
  Daniel Hiltgen a2ad952440 Fix windows system memory lookup 1 سال پیش
  Daniel Hiltgen d966b730ac Switch windows build to fully dynamic 1 سال پیش
  Daniel Hiltgen 7555ea44f8 Revamp the dynamic library shim 1 سال پیش
  Daniel Hiltgen 6558f94ed0 Fix darwin intel build 1 سال پیش
  Daniel Hiltgen 35934b2e05 Adapted rocm support to cgo based llama.cpp 1 سال پیش