Commits

Branch selector

 master

User selector

slaren

Datepicker

All time

Commit History

Commits on Sep 30, 2023

• ggml-cuda : perform cublas mat mul of quantized types as f16 (ggerganov#3412)

  slaren

  committedSep 30, 2023
  f5ef5cf

Commits on Sep 29, 2023

• llama.cpp : add documentation about rope_freq_base and scale values (ggerganov#3401)

  slaren

  committedSep 29, 2023
  40e07a6

Commits on Sep 28, 2023

• llama.cpp : split llama_context_params into model and context params (ggerganov#3301)

  slaren

  committedSep 28, 2023
  16bc66d

• ggml-cuda : perform cublas fp16 matrix multiplication as fp16 (ggerganov#3370)

  slaren

  committedSep 28, 2023
  da04003

Commits on Sep 23, 2023

• llama-bench : add README (ggerganov#3317)

  slaren

  committedSep 23, 2023
  c091cdf

Commits on Sep 17, 2023

• llama.cpp : show model size and BPW on load (ggerganov#3223)

  slaren

  committedSep 17, 2023
  8b428c9

Commits on Sep 8, 2023

• ggml-alloc : correctly check mmap return value for errors (ggerganov#3075)

  slaren

  committedSep 8, 2023
  ebc9608

Commits on Sep 7, 2023

• llama-bench : use two tokens in the warmup run for prompt evals (ggerganov#3059)

  slaren

  committedSep 7, 2023
  15b67a6

Commits on Sep 4, 2023

• ggml-opencl : store GPU buffer in ggml_tensor::extra (ggerganov#2994)

  slaren

  committedSep 4, 2023
  bd33e5a

Commits on Sep 3, 2023

• ggml-alloc : use virtual memory for measurement (ggerganov#2973)

  slaren

  committedSep 3, 2023
  cf9b084

Commits on Aug 31, 2023

• remove convert-llama-7b-pth-to-gguf.py and convert-llama-hf-to-gguf.py (ggerganov#2906)

  slaren

  committedAug 31, 2023
  528134d

Commits on Aug 29, 2023

• ggml : add view_src and view_offs to ggml_tensor for views (ggerganov#2874)

  slaren

  committedAug 29, 2023
  06abf8e

• remove outdated references to -eps and -gqa from README (ggerganov#2881)

  slaren

  committedAug 29, 2023
  c03a243

Commits on Aug 28, 2023

• llama-bench : set locale to utf8 (ggerganov#2832)

  slaren

  committedAug 28, 2023
  43033b7

Commits on Aug 27, 2023

• ci : add LoRA test to CI (ggerganov#2650)

  

  slaren

  and

  ggerganov

  committedAug 27, 2023
  789c8c9

Commits on Aug 25, 2023

• llama-bench : add model sizes (ggerganov#2771)

  slaren

  committedAug 25, 2023
  154725c

• convert.py : export rope freq_base when converting CodeLlama from an HF model (ggerganov#2773)

  slaren

  committedAug 25, 2023
  12e2e33

Commits on Aug 24, 2023

• convert.py : try to determine n_ctx automatically for CodeLlama (ggerganov#2770)

  slaren

  committedAug 24, 2023
  d0f77b1

• gguf : add rope_freq_base parameter for CodeLlama (ggerganov#2769)

  slaren

  committedAug 24, 2023
  0d3094f

• fix convert.py for codellama, add llama 34B to the list of recognized models (ggerganov#2768)

  slaren

  committedAug 24, 2023
  fea95c6

Commits on Aug 23, 2023

• fix convert-lora-to-ggml.py (ggerganov#2738)

  slaren

  committedAug 23, 2023
  335acd2

Commits on Aug 22, 2023

• embedding : evaluate prompt in batches (ggerganov#2713)

  slaren

  committedAug 22, 2023
  519c981

• ggml-cuda : use graph allocator (ggerganov#2684)

  

  slaren

  and

  JohannesGaessler

  committedAug 22, 2023
  1123f7f

• llama-bench : minor fixes (ggerganov#2695)

  slaren

  committedAug 22, 2023
  8e4364f

Commits on Aug 20, 2023

• ggml : move all type info to ggml_type_traits (ggerganov#2663)

  slaren

  committedAug 20, 2023
  9e232f0

Commits on Aug 18, 2023

• llama : add benchmark example (ggerganov#2626)

  

  slaren

  and

  JohannesGaessler

  committedAug 18, 2023
  097e121

Commits on Jul 31, 2023

• Fix Metal backend broken from the allocator changes (ggerganov#2455)

  slaren

  committedJul 31, 2023
  9d2382b

Commits on Jul 30, 2023

• ggml : add graph tensor allocator (ggerganov#2411)

  slaren

  committedJul 30, 2023
  a113689

Commits on Jul 26, 2023

• ggml : fix assert in ggml_set_unary_op (ggerganov#2410)

  slaren

  committedJul 26, 2023
  b5472ea

• ggml : allocate graphs in a context (ggerganov#2392)

  

  slaren

  and

  ggerganov

  committedJul 26, 2023
  5488fb7

Commits on Jul 25, 2023

• ggml : fix ggml_flash_attn to use op_params (ggerganov#2387)

  slaren

  committedJul 25, 2023
  07aaa0f

• ggml : improve graph build time via hash table lookup (ggerganov#2329)

  slaren

  committedJul 25, 2023
  da18898

• server: add rms_norm_eps parameter (ggerganov#2380)

  slaren

  committedJul 25, 2023
  d5512b7

Commits on Jul 24, 2023

• make rms_norm_eps a parameter (ggerganov#2374)

  slaren

  committedJul 24, 2023
  41c6741

Commits on Jul 23, 2023

• fix n_tasks (ggerganov#2342)

  slaren

  committedJul 23, 2023
  3602ac4

Pagination

PreviousNext