Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 18, 2025

• Adding linenoise.cpp to llama-run (#11252)

  Show description for a1649cc
  ericcurtin

  authored
  a1649cc

  Copy full SHA for a1649cc

• cmake : add sanitizer flags for llama.cpp (#11279)

  Show description for 4dd34ff
  

  ggerganov

  and

  JohannesGaessler

  authored
  4dd34ff

  Copy full SHA for 4dd34ff

• server : implement cancellable request (#11285)

  Show description for f30f099
  ngxson

  authored
  f30f099

  Copy full SHA for f30f099

• scripts : restore hf.sh (#11288)

  Show description for f26c874
  ggerganov

  authored
  f26c874

  Copy full SHA for f26c874

• tts : add guide tokens support (#11186)

  Show description for 6390a99
  LostRuins

  authored
  6390a99

  Copy full SHA for 6390a99

• vulkan: fix coopmat2 flash attention for non-contiguous inputs (#11281)

  Show description for 44e18ef
  jeffbolznv

  authored
  44e18ef

  Copy full SHA for 44e18ef

Commits on Jan 17, 2025

• llama.android: add field formatChat to control whether to parse special tokens when send message (#11270)

  codezjx

  authored
  3edfa7d

  Copy full SHA for 3edfa7d

• rpc : early register backend devices (#11262)

  Show description for 667d728
  rgerganov

  authored
  667d728

  Copy full SHA for 667d728

• vocab : fix double-eos check (#11273)

  Show description for a133566
  ggerganov

  authored
  a133566

  Copy full SHA for a133566

• llama : fix deprecation message: vocabable -> vocab (#11269)

  dwrensha

  authored
  960ec65

  Copy full SHA for 960ec65

• README : added kalavai to infrastructure list (#11216)

  musoles

  authored
  7a689c4

  Copy full SHA for 7a689c4

Commits on Jan 16, 2025

• vulkan: support copy from f32 to q4_0/q4_1/q5_0/q5_1/q8_0/iq4_nl (#11166)

  Show description for bd38dde
  jeffbolznv

  authored
  bd38dde

  Copy full SHA for bd38dde

• vulkan: optimize coopmat2 q4_k/q5_k dequant functions. (#11206)

  Show description for 466300f
  jeffbolznv

  authored
  466300f

  Copy full SHA for 466300f

• vulkan: optimize coopmat2 q2_k dequant function (#11130)

  jeffbolznv

  authored
  206bc53

  Copy full SHA for 206bc53

• llama : add internlm3 support (#11233)

  Show description for 4dbc8b9
  RunningLeon

  authored
  4dbc8b9

  Copy full SHA for 4dbc8b9

• CUDA: backwards pass for misc. ops, add tests (#11257)

  Show description for 9c8dcef
  JohannesGaessler

  authored
  9c8dcef

  Copy full SHA for 9c8dcef

• llama : add `llama_model_load_from_splits` (#11255)

  Show description for 681149c
  ngxson

  authored
  681149c

  Copy full SHA for 681149c

• ggml: aarch64: implement SVE kernels for q4_K_q8_K vector dot (#11227)

  Show description for c67cc98
  

  fj-y-saito

  and

  ggerganov

  authored
  c67cc98

  Copy full SHA for c67cc98

Commits on Jan 15, 2025

• vulkan: scale caching for k quants + misc fixes (#11081)

  Show description for adc5dd9
  netrunnereve

  authored
  adc5dd9

  Copy full SHA for adc5dd9

• ci : use -no-cnv in gguf-split tests (#11254)

  Show description for f11cfdf
  ggerganov

  authored
  f11cfdf

  Copy full SHA for f11cfdf

• fix: ggml: fix vulkan-shaders-gen build (#10448)

  Show description for 1d85043
  sparkleholic

  authored
  1d85043

  Copy full SHA for 1d85043

• RoPE: fix back, CUDA support for back + noncont. (#11240)

  Show description for 432df2d
  JohannesGaessler

  authored
  432df2d

  Copy full SHA for 432df2d

• examples : add embd_to_audio to tts-outetts.py [no ci] (#11235)

  Show description for 0ccd7f3
  danbev

  authored
  0ccd7f3

  Copy full SHA for 0ccd7f3

• SYCL: Add gated linear attention kernel (#11175)

  Show description for f446c2c
  qnixsynapse

  authored
  f446c2c

  Copy full SHA for f446c2c

Commits on Jan 14, 2025

• ci : add -no-cnv for tests (#11238)

  ngxson

  authored
  b4d92a5

  Copy full SHA for b4d92a5

• vocab : add dummy tokens for "no_vocab" type (#11231)

  Show description for bbf3e55
  ggerganov

  authored
  bbf3e55

  Copy full SHA for bbf3e55

• server : Improve code snippets direction between RTL text (#11221)

  ebraminio

  authored
  c5bf0d1

  Copy full SHA for c5bf0d1

• Refactor test-chat-template.cpp (#11224)

  Show description for 091592d
  ochafik

  authored
  091592d

  Copy full SHA for 091592d

• sync : ggml

  ggerganov

  committed
  44d1e79

  Copy full SHA for 44d1e79

• scripts : sync gguf (cont)

  ggerganov

  committed
  a4f3f5d

  Copy full SHA for a4f3f5d

• scripts : sync gguf

  ggerganov

  committed
  48e1ae0

  Copy full SHA for 48e1ae0

• scripts : sync opencl

  ggerganov

  committed
  d00a80e

  Copy full SHA for d00a80e

Commits on Jan 13, 2025

• server : (UI) Improve messages bubble shape in RTL (#11220)

  Show description for 504af20
  ebraminio

  authored
  504af20

  Copy full SHA for 504af20

• cli : auto activate conversation mode if chat template is available (#11214)

  Show description for 84a4481
  ngxson

  authored
  84a4481

  Copy full SHA for 84a4481

• cuda : CUDA Graph Compute Function Refactor (precursor for performance improvements) (#11042)

  Show description for 39509fb
  

  aendk

  and

  slaren

  authored
  39509fb

  Copy full SHA for 39509fb

Pagination

Previous

Next