Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jun 15, 2024

• Add cvector-generator example (ggerganov#7514)

  

  ngxson

  and

  christianazinn

  authored
  0c7b359
  More actions

• [SYCL] remove global variables (ggerganov#7710)

  airMeng

  authored
  7b2f4a7
  More actions

Commits on Jun 14, 2024

• ci : fix macos x86 build (ggerganov#7940)

  olexiyb

  authored
  f8ec887
  More actions

• CUDA: faster q2_K, q3_K MMQ + int8 tensor cores (ggerganov#7921)

  JohannesGaessler

  authored
  76d66ee
  More actions

• metal : utilize max shared memory for mul_mat_id (ggerganov#7935)

  ggerganov

  authored
  66ef1ce
  More actions

• llama-bench : fix RPC indication (ggerganov#7936)

  rgerganov

  authored
  e65bbf6
  More actions

• llama : more checks before assuming FIM tokens (ggerganov#7644)

  CISC

  authored
  6fcd133
  More actions

• convert : add Poro-34B-chat tokenizer support (ggerganov#7713)

  

  ezosa

  and

  ggerganov

  authored
  41b9260
  More actions

Commits on Jun 13, 2024

• rpc : fix ggml_backend_rpc_supports_buft() (ggerganov#7918)

  rgerganov

  authored
  172c825
  More actions

• readme : Remove outdated instructions from README.md (ggerganov#7914) [no ci]

  Galunid

  authored
  a55eb1b
  More actions

• move BLAS to a separate backend (ggerganov#6210)

  

  slaren

  and

  ggerganov

  authored
  f578b86
  More actions

Commits on Jun 12, 2024

• build: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (ggerganov#7809)

  

  ochafik

  and

  HanClinto

  authored
  1c641e6
  More actions

• CUDA: fix broken oob check for FA vec f32 kernel (ggerganov#7904)

  JohannesGaessler

  authored
  9635529
  More actions

• tests : add non-cont unary tests (ggerganov#7857)

  ggerganov

  authored
  a9cae48
  More actions

• ggml : improve ggml_is_contiguous logic (ggerganov#7856)

  ggerganov

  authored
  bfaa676
  More actions

• server : restore numeric prompts (ggerganov#7883)

  ggerganov

  authored
  704a35b
  More actions

• update intel docker oneapi-basekit to 2024.1.1-devel-ubuntu22.04 (ggerganov#7894)

  airMeng

  authored
  dcf7527
  More actions

• Fix a typo and add Fedora 40 pacakge to install for Vulkan (ggerganov#7794) [no ci]

  metal3d

  authored
  f2b5764
  More actions

Commits on Jun 11, 2024

• vulkan: select only one device for single gpu with multiple drivers (ggerganov#7582)

  Adriankhl

  authored
  73bac2b
  More actions

• Update Vulkan RoPE implementation (ggerganov#7818)

  

  0cc4m

  and

  slaren

  authored
  ef52d1d
  More actions

• fix broken link in pr template (ggerganov#7880) [no ci]

  

  deven367

  and

  mofosyne

  authored
  14f8352
  More actions

• github: move PR template to .github/ root (ggerganov#7868)

  mofosyne

  authored
  6fe42d0
  More actions

• llama-bench: more compact markdown tables (ggerganov#7879)

  JohannesGaessler

  authored
  148995e
  More actions

• tests : check the Python version (ggerganov#7872)

  ggerganov

  authored
  4bfe50f
  More actions

• CUDA: int8 tensor cores for MMQ (q4_K, q5_K, q6_K) (ggerganov#7860)

  JohannesGaessler

  authored
  bdcb8f4
  More actions

• fix CUDA CI by using a windows-2019 image (ggerganov#7861)

  slaren

  authored
  c2ce6c4
  More actions

• json: refine constraint for whitespace to avoid runaways yet allow pretty print (ggerganov#7866)

  ochafik

  authored
  b61eb96
  More actions

• json: document schema conversion in GBNF readme, align manual grammar examples & converters (ggerganov#7841)

  ochafik

  authored
  396b18d
  More actions

Commits on Jun 10, 2024

• cmake : fix CMake requirement for CUDA (ggerganov#7821)

  cebtenzzre

  authored
  864a99e
  More actions

• ci : try win-2019 on server windows test (ggerganov#7854)

  slaren

  authored
  fd5ea0f
  More actions

• examples : remove --instruct remnants (ggerganov#7846)

  ggerganov

  authored
  c28a839
  More actions

• server : improve "prompt" handling (ggerganov#7847)

  ggerganov

  authored
  d9da0e4
  More actions

• CUDA: use tensor cores for MMQ (ggerganov#7676)

  JohannesGaessler

  authored
  1f0dabd
  More actions

• use the correct SYCL context for host USM allocations (ggerganov#7777)

  bashbaug

  authored
  af4ae50
  More actions

Commits on Jun 9, 2024

• flake.lock: Update (ggerganov#7838)

  

  ggerganov

  and

  github-actions[bot]

  authored
  10ceba3
  More actions

Pagination

PreviousNext