Commit History

Автор SHA1 Съобщение Дата
  Georgi Gerganov a4090d1174 llama : remove llama_kv_cache_view API + remove deprecated (#13653) преди 8 месеца
  Georgi Gerganov e298d2fbd0 kv-cache : add SWA support (#13194) преди 8 месеца
  psocolovsky 1dfbf2cf3a common : add load_progress_callback (#13617) преди 8 месеца
  Isaac McFadyen 6a2bc8bfb7 server : added --no-prefill-assistant flag (#13608) преди 8 месеца
  Olivier Chafik 3198405e98 `common`: add partial regex support (#12808) преди 8 месеца
  Johannes Gäßler 10d2af0eaa llama/ggml: add LLM training support (#10544) преди 8 месеца
  David Huang 7f323a589f Add `--no-op-offload` to improve `-ot` pp perf in MoE models like llama4 400B (#13386) преди 8 месеца
  Bartowski efb8b47eda imatrix : Add --parse-special for enabling parsing of special tokens in imatrix calculation (#13389) преди 8 месеца
  Georgi Gerganov 51fb96b1ff context : remove logits_all flag (#13284) преди 8 месеца
  Georgi Gerganov 4773d7a02f examples : remove infill (#13283) преди 8 месеца
  oobabooga 233461f812 sampling : Integrate Top-nσ into main sampling chain (and add it to the server) (#13264) преди 8 месеца
  Xuan-Son Nguyen 9b61acf060 mtmd : rename llava directory to mtmd (#13311) преди 8 месеца
  Diego Devesa 1d36b3670b llama : move end-user examples to tools directory (#13249) преди 8 месеца
  Xuan-Son Nguyen 7c727fbe39 arg : add --no-mmproj-offload (#13093) преди 9 месеца
  Xuan-Son Nguyen 80982e815e arg : clean up handling --mmproj with -hf (#13082) преди 9 месеца
  tastelikefeet b2034c2b55 contrib: support modelscope community (#12664) преди 9 месеца
  Diego Devesa e0e912f49b llama : add option to override model tensor buffers (#11397) преди 9 месеца
  Xuan-Son Nguyen 42eb248f46 common : remove json.hpp from common.cpp (#12697) преди 9 месеца
  Xuan-Son Nguyen 267c1399f1 common : refactor downloading system, handle mmproj with -hf option (#12694) преди 9 месеца
  marcoStocchi 6ef79a67ca common : refactor '-o' option (#12278) преди 10 месеца
  Olivier Chafik 669912d9a5 `tool-call`: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammars (#12034) преди 10 месеца
  Sigbjørn Skjæret 56d7a9f812 main: allow preloading conversation with -p and add -st / --single-turn (#12145) преди 10 месеца
  dm4 c43af9276b tts: add speaker file support (#12048) преди 10 месеца
  Sigbjørn Skjæret 45a8e76745 common : add --system-prompt parameter, replace behavior of -p in conversation mode (#12131) преди 10 месеца
  Georgi Gerganov abd4d0bc4f speculative : update default params (#11954) преди 11 месеца
  Olivier Chafik 63e489c025 tool-call: refactor common chat / tool-call api (+ tests / fixes) (#11900) преди 11 месеца
  Daniel Bevenius c48f630d1c llama : add --completion-bash option (#11846) преди 11 месеца
  Olivier Chafik c7f460ab88 `server`: fix tool-call of DeepSeek R1 Qwen, return reasoning_content (Command 7RB & DeepSeek R1) unless `--reasoning-format none` (#11607) преди 11 месеца
  Vinesh Janarthanan 27e8a23300 sampling: add Top-nσ sampler (#11223) преди 11 месеца
  bandoti fef0cbeadf cleanup: fix compile warnings associated with gnu_printf (#11811) преди 11 месеца