コミット履歴

作者 SHA1 メッセージ 日付
  Daniel Hiltgen 3518aaef33 Merge pull request #4218 from dhiltgen/auto_parallel 10 ヶ月 前
  Blake Mizerany cb42e607c5 llm: speed up gguf decoding by a lot (#5246) 10 ヶ月 前
  Daniel Hiltgen 9929751cc8 Disable concurrency for AMD + Windows 10 ヶ月 前
  Daniel Hiltgen 17b7186cd7 Enable concurrency by default 1 年間 前
  Daniel Hiltgen 6f351bf586 review comments and coverage 11 ヶ月 前
  Daniel Hiltgen ff4f0cbd1d Prevent multiple concurrent loads on the same gpus 11 ヶ月 前
  Daniel Hiltgen fc37c192ae Refine CPU load behavior with system memory visibility 11 ヶ月 前
  Daniel Hiltgen 434dfe30c5 Reintroduce nvidia nvml library for windows 11 ヶ月 前
  Daniel Hiltgen 48702dd149 Harden unload for empty runners 11 ヶ月 前
  Daniel Hiltgen 5e8ff556cb Support forced spreading for multi GPU 1 年間 前
  Michael Yang e40145a39d lint 11 ヶ月 前
  Michael Yang c895a7d13f some gocritic 11 ヶ月 前
  Michael Yang 04f3c12bb7 replace x/exp/slices with slices 11 ヶ月 前
  Patrick Devine 4cc3be3035 Move envconfig and consolidate env vars (#4608) 11 ヶ月 前
  Sang Park 4434d7f447 Correct typo in error message (#4535) 11 ヶ月 前
  Daniel Hiltgen ec231a7923 Remove VRAM convergence check for windows 11 ヶ月 前
  Patrick Devine 6845988807 Ollama `ps` command for showing currently loaded models (#4327) 11 ヶ月 前
  Daniel Hiltgen 4142c3ef7c Always use the sorted list of GPUs 11 ヶ月 前
  Jeffrey Morgan bb6fd02298 Don't clamp ctx size in `PredictServerFit` (#4317) 11 ヶ月 前
  Daniel Hiltgen 354ad9254e Wait for GPU free memory reporting to converge 11 ヶ月 前
  Jeffrey Morgan c9f98622b1 Skip scheduling cancelled requests, always reload unloaded runners (#4189) 1 年間 前
  Jeffrey Morgan dfa2f32ca0 unload in critical section (#4187) 1 年間 前
  Daniel Hiltgen f56aa20014 Centralize server config handling 1 年間 前
  Jeffrey Morgan 942c979232 allocate a large enough kv cache for all parallel requests (#4162) 1 年間 前
  Daniel Hiltgen 20f6c06569 Make maximum pending request configurable 1 年間 前
  Mark Ward 63c763685f log when the waiting for the process to stop to help debug when other tasks execute during this wait. 1 年間 前
  Mark Ward f4a73d57a4 fix runner expire during active use. Clearing the expire timer as it is used. Allowing the finish to assign an expire timer so that the runner will expire after no use. 1 年間 前
  Daniel Hiltgen d6e3b64582 Fix concurrency for CPU mode 1 年間 前
  Daniel Hiltgen 9b5a3c5991 Merge pull request #3914 from dhiltgen/mac_perf 1 年間 前
  Jeffrey Morgan 00b0699c75 Reload model if `num_gpu` changes (#3920) 1 年間 前