Commits

Branch selector

master

User selector

All users

Datepicker

All time

Commit history

Commits on Aug 8, 2026

• CUDA: fuse rms_norm + mul + rope (+ view + set_rows) (#26767)

  Show description for 687e778
  grafail

  authored
  687e778View commit details

  Copy full SHA for 687e778

  Browse repository at this point

• server, ui: only offer a working directory when a tool reads it (#26762)

  Show description for 18f7ad7
  ServeurpersoCom

  authored
  18f7ad7View commit details

  Copy full SHA for 18f7ad7

  Browse repository at this point

• server: add initial tool isolation support (via docker) (#26507)

  Show description for dd2c7c4
  

  ngxson

  and

  ServeurpersoCom

  authored
  dd2c7c4View commit details

  Copy full SHA for dd2c7c4

  Browse repository at this point

• CUDA: fix thread/block count in quantized cpy kernel launches (#26731)

  Show description for 69bf643
  grafail

  authored
  69bf643View commit details

  Copy full SHA for 69bf643

  Browse repository at this point

Commits on Aug 7, 2026

• tts: account for the vocoder pass in the timings line (#26733)

  Show description for 3653e6d
  ServeurpersoCom

  authored
  3653e6dView commit details

  Copy full SHA for 3653e6d

  Browse repository at this point

• allozaur/feat/chat form contenteditable (#26717)

  Show description for fc6545d
  ServeurpersoCom

  authored
  fc6545dView commit details

  Copy full SHA for fc6545d

  Browse repository at this point

• tests : speed-up server test suite 3x (#26734)

  Show description for 1621a3d
  ggerganov

  authored
  1621a3dView commit details

  Copy full SHA for 1621a3d

  Browse repository at this point

• allozaur/feat/chat slash commands (#26716)

  Show description for 6de1b63
  ServeurpersoCom

  authored
  6de1b63View commit details

  Copy full SHA for 6de1b63

  Browse repository at this point

• sycl: coalesce the ssm_conv window loads (#26612)

  Show description for f8e3026
  Titaniumtown

  authored
  f8e3026View commit details

  Copy full SHA for f8e3026

  Browse repository at this point

• metal : fix NORM/RMS_NORM for row lengths that leave a partial simdgroup (#26708)

  Show description for a194a75
  robertomeroni

  authored
  a194a75View commit details

  Copy full SHA for a194a75

  Browse repository at this point

• ui: Filesystem `@mentions` for Chat Form (#26715)

  Show description for 2363478
  ServeurpersoCom

  authored
  2363478View commit details

  Copy full SHA for 2363478

  Browse repository at this point

• mtmd: fix longest_edge ignoring min/max pixels (#26638)

  Show description for 4cb22cd
  ngxson

  authored
  4cb22cdView commit details

  Copy full SHA for 4cb22cd

  Browse repository at this point

• sync : ggml

  ggerganov

  committed
  4cf5cabView commit details

  Copy full SHA for 4cf5cab

  Browse repository at this point

• ggml : bump version to 0.19.0 (ggml/1581)

  ggerganov

  committed
  933f46fView commit details

  Copy full SHA for 933f46f

  Browse repository at this point

• server : clarify comment in eval_llama_cmpl_schema [no ci] [no release] (#26720)

  danbev

  authored
  9ba73fdView commit details

  Copy full SHA for 9ba73fd

  Browse repository at this point

• webui: load the model selected via ?model= when ?load=true (#26707)

  Show description for f4f7758
  erusev

  authored
  f4f7758View commit details

  Copy full SHA for f4f7758

  Browse repository at this point

• ui: set npm `min-release-age` to protect against supply-chain attacks (#26711)

  Show description for 34e9ee5
  nikwen

  authored
  34e9ee5View commit details

  Copy full SHA for 34e9ee5

  Browse repository at this point

• server: (router) add LRU scheduler (#26572)

  Show description for dff15d4
  ngxson

  authored
  dff15d4View commit details

  Copy full SHA for dff15d4

  Browse repository at this point

• server: (router) do not evict busy models (#26567)

  ngxson

  authored
  e1470eeView commit details

  Copy full SHA for e1470ee

  Browse repository at this point

• mtmd: stop feeding the text stream again during Qwen3-TTS generation (#26706)

  Show description for 217df17
  ServeurpersoCom

  authored
  217df17View commit details

  Copy full SHA for 217df17

  Browse repository at this point

• ggml : add aarch64 HWCAP fallbacks and fix fp16 variant detection (#25554)

  Show description for cb26014
  kilian-hu

  authored
  cb26014View commit details

  Copy full SHA for cb26014

  Browse repository at this point

• ui: read model modalities from the router model list (#26709)

  Show description for 82bb485
  ServeurpersoCom

  authored
  82bb485View commit details

  Copy full SHA for 82bb485

  Browse repository at this point

• Mitigate crashing issue on Windows MSYS2 UCRT64 environment (GCC 16.1.0) (#26555)

  rillomas

  authored
  42e9881View commit details

  Copy full SHA for 42e9881

  Browse repository at this point

• sycl: fix UE4M3 parsing (#25608)

  Show description for fc3f10b
  clee

  authored
  fc3f10bView commit details

  Copy full SHA for fc3f10b

  Browse repository at this point

• sycl: *glu flat path (#26354)

  Show description for 6b5c2ef
  Titaniumtown

  authored
  6b5c2efView commit details

  Copy full SHA for 6b5c2ef

  Browse repository at this point

• sycl : Support DSv4 OPs: LIGHTNING_INDEXER,DSV4_HC_COMB,DSV4_HC_POST,DSV4_HC_PRE (#26568)

  Show description for 31558db
  arthw

  authored
  31558dbView commit details

  Copy full SHA for 31558db

  Browse repository at this point

• sycl : update guide Q&A and script for device setting (#26442)

  arthw

  authored
  c1f4109View commit details

  Copy full SHA for c1f4109

  Browse repository at this point

• sycl : fix error Error OP FLASH_ATTN_EXT on arc770 (#26441)

  arthw

  authored
  eef5f3eView commit details

  Copy full SHA for eef5f3e

  Browse repository at this point

• sycl : enhance OP set_rows to support all missed data types (#26515)

  Show description for c074cb3
  arthw

  authored
  c074cb3View commit details

  Copy full SHA for c074cb3

  Browse repository at this point

• cuda: fix warnings for unused variable/function (#26688)

  dfriehs

  authored
  5b87ed3View commit details

  Copy full SHA for 5b87ed3

  Browse repository at this point

• ci: abort if build requirements are missing (#26368)

  Show description for d8d9887
  nikwen

  authored
  d8d9887View commit details

  Copy full SHA for d8d9887

  Browse repository at this point

• metal : avoid `threadgroup` matrix array instantiation in kernel_lightning_indexer (#26646)

  Show description for e40bf88
  JamePeng

  authored
  e40bf88View commit details

  Copy full SHA for e40bf88

  Browse repository at this point

Commits on Aug 6, 2026

• mtmd: add chunk save/load function (#26645)

  Show description for 15586e2
  ngxson

  authored
  15586e2View commit details

  Copy full SHA for 15586e2

  Browse repository at this point

• server: fix empty response for /cors-proxy (#26656)

  ngxson

  authored
  6a32c29View commit details

  Copy full SHA for 6a32c29

  Browse repository at this point

• convert : fix DeepseekV4 rope parameters with transformers 5.x (#26673)

  CISC

  authored
  eb5667aView commit details

  Copy full SHA for eb5667a

  Browse repository at this point