Code

Gg/qwen2 f32 prec#228

Code

Merged

Nexesenex merged 14 commits intoNexesenex:qwenfixNexesenex/croco.cpp:qwenfixfrom ggerganov:gg/qwen2-f32-precggerganov/llama.cpp:gg/qwen2-f32-precCopy head branch name to clipboardJul 11, 2024

+2,793-99

Commits

Commits on Jul 9, 2024

• sycl : Reenabled mmvq path for the SYCL Nvidia Backend (#8372)

  Show description for 5b0b8d8
  Alcpz

  authored
  5b0b8d8

  Copy full SHA for 5b0b8d8

• make/cmake: LLAMA_NO_CCACHE -> GGML_NO_CCACHE (#8392)

  JohannesGaessler

  authored
  a03e8dd

  Copy full SHA for a03e8dd

• Deprecation warning to assist with migration to new binary names (#8283)

  Show description for e500d61
  HanClinto

  authored
  e500d61

  Copy full SHA for e500d61

• Update README.md to fix broken link to docs (#8399)

  Show description for fd560fe
  andysalerno

  authored
  fd560fe

  Copy full SHA for fd560fe

• Server: Enable setting default sampling parameters via command-line (#8402)

  Show description for a59f8fd
  HanClinto

  authored
  a59f8fd

  Copy full SHA for a59f8fd

Commits on Jul 10, 2024

• py : fix extra space in convert_hf_to_gguf.py (#8407)

  laik

  authored
  8f0fad4

  Copy full SHA for 8f0fad4

• py : fix converter for internlm2 (#8321)

  Show description for e4dd31f
  RunningLeon

  authored
  e4dd31f

  Copy full SHA for e4dd31f

• llama : add assert about missing llama_encode() call (#8400)

  Show description for a8be1e6
  

  fairydreaming

  and

  sszymczy

  authored
  a8be1e6

  Copy full SHA for a8be1e6

• msvc : silence codecvt c++17 deprecation warnings (#8395)

  iboB

  authored
  7a80710

  Copy full SHA for 7a80710

• llama : C++20 compatibility for u8 strings (#8408)

  iboB

  authored
  cc61948

  Copy full SHA for cc61948

• gguf-py rel pipeline (#8410)

  Show description for 83321c6
  monatis

  authored
  83321c6

  Copy full SHA for 83321c6

• ggml : add AArch64 optimized GEMV and GEMM Q4 kernels (#5780)

  Show description for 0f1a39f
  Dibakar

  authored
  0f1a39f

  Copy full SHA for 0f1a39f

• ggml : move sgemm sources to llamafile subfolder (#8394)

  Show description for 6b2a849
  ggerganov

  authored
  6b2a849

  Copy full SHA for 6b2a849

• llama : use F32 precision in Qwen2 attention and no FA

  ggerganov

  committed
  7c9e9a2

  Copy full SHA for 7c9e9a2