Commits

Branch selector

 grammar

User selector

All users

Datepicker

All time

Commit History

Commits on Jun 13, 2023

• WIP

  xaptronic

  committed
  f2ed4ba

  Copy full SHA for f2ed4ba

Commits on Jun 12, 2023

• add missing include

  ejones

  committed
  3e78f00

  Copy full SHA for 3e78f00

• Merge remote-tracking branch 'refs/remotes/upstream/master' into grammar

  ejones

  committed
  56904ca

  Copy full SHA for 56904ca

• add comments to grammar syntax and allow newlines where unambiguous

  ejones

  committed
  98a9587

  Copy full SHA for 98a9587

• handle & print parser errors

  ejones

  committed
  674bb08

  Copy full SHA for 674bb08

• fix whitespace errors

  ejones

  committed
  9e77f42

  Copy full SHA for 9e77f42

• allow loading grammar from file

  ejones

  committed
  834d423

  Copy full SHA for 834d423

Commits on Jun 11, 2023

• Fix issue where interactive mode crashes when input exceeds ctx size (#1789)

  Show description for fa84c4b
  KerfuffleV2

  authored
  fa84c4b

  Copy full SHA for fa84c4b

• Fixed WSL cuda's OOM error (#1594)

  Show description for 12b063f
  

  JoelSeniorLiang

  and

  liang

  authored
  12b063f

  Copy full SHA for 12b063f

• Update SHA256SUMS with current hashes for models quantized using q4_0 (#1798)

  rlanday

  authored
  31d2b5f

  Copy full SHA for 31d2b5f

Commits on Jun 10, 2023

• cmake : fix Metal build (close #1791)

  ggerganov

  committed
  4de0334

  Copy full SHA for 4de0334

• k-quants : GCC12 compilation fix (#1792)

  vagran

  authored
  3f12231

  Copy full SHA for 3f12231

• metal : fix issue with ggml-metal.metal path. Closes #1769 (#1782)

  Show description for 303f580
  abetlen

  authored
  303f580

  Copy full SHA for 303f580

• doc : fix wrong address of BLIS.md (#1772)

  Show description for 059e990
  Aisuko

  authored
  059e990

  Copy full SHA for 059e990

• ggml : force no_alloc == false when creating opt tensors (close #1699)

  Show description for 17c10ac
  ggerganov

  committed
  17c10ac

  Copy full SHA for 17c10ac

• metal : add Q4_1 implementation (#1785)

  Show description for e9b66ee
  

  ikawrakow

  and

  Kawrakow

  authored
  e9b66ee

  Copy full SHA for e9b66ee

• llama : support requantizing models instead of only allowing quantization from 16/32bit (#1691)

  Show description for 4f0154b
  KerfuffleV2

  authored
  4f0154b

  Copy full SHA for 4f0154b

• ggml : workaround for missing _mm256_setr_m128i in GCC < 8 (#1638)

  xingchensong

  authored
  ef3171d

  Copy full SHA for ef3171d

• make : add SSSE3 compilation use case (#1659)

  

  rankaiyx

  authored
  555275a

  Copy full SHA for 555275a

Commits on Jun 9, 2023

• OpenCL: Add release memory (#1741)

  Show description for 98ed165
  edp1096

  authored
  98ed165

  Copy full SHA for 98ed165

• Windows nvcc workaround (#1753)

  Show description for ae9663f
  JohannesGaessler

  authored
  ae9663f

  Copy full SHA for ae9663f

• metal : fix build "tanhf" -> "tanh"

  ggerganov

  committed
  b33dee2

  Copy full SHA for b33dee2

• metal : add GELU implementation (#1770)

  Show description for 92f44ff
  manyoso

  authored
  92f44ff

  Copy full SHA for 92f44ff

• metal : faster q4_0 (#1775)

  Show description for 245fc3c
  

  ikawrakow

  and

  Kawrakow

  authored
  245fc3c

  Copy full SHA for 245fc3c

• llama, main : constrain sampling to grammar

  ejones

  committed
  fd0eb66

  Copy full SHA for fd0eb66

Commits on Jun 8, 2023

• metal : add Q2_K implementation (#1762)

  Show description for 72ff528
  

  ikawrakow

  and

  Kawrakow

  authored
  72ff528

  Copy full SHA for 72ff528

• Revert "ggml : load data into int8x16x4_t using vld4q_s8 on arm64 (#1738)"

  Show description for 0bf7cf1
  ggerganov

  committed
  0bf7cf1

  Copy full SHA for 0bf7cf1

• ggml : load data into int8x16x4_t using vld4q_s8 on arm64 (#1738)

  lindeer

  authored
  8432d4d

  Copy full SHA for 8432d4d

• metal : Q6_K implementation (#1752)

  Show description for 0f291e1
  

  ikawrakow

  and

  Kawrakow

  authored
  0f291e1

  Copy full SHA for 0f291e1

• Add llama.cpp docker support for non-latin languages (#1673)

  Show description for 8fc8179
  qingfengfenga

  authored
  8fc8179

  Copy full SHA for 8fc8179

• ggml : fix fprintf warnings (#1720)

  sroussey

  authored
  b50b570

  Copy full SHA for b50b570

• clang-tidy : restore dot file from accidental deletion

  ggerganov

  committed
  53aba3f

  Copy full SHA for 53aba3f

• metal : add Q4_K implementation (#1733)

  Show description for 4161bdc
  

  ikawrakow

  and

  Kawrakow

  authored
  4161bdc

  Copy full SHA for 4161bdc

• k-quants : add missing compile definition to CMakeLists (#1748)

  johnson442

  authored
  0035858

  Copy full SHA for 0035858

Commits on Jun 7, 2023

• k-quants : allow to optionally disable at compile time (#1734)

  Show description for 5c64a09
  ggerganov

  authored
  5c64a09

  Copy full SHA for 5c64a09

Pagination

Previous

Next