Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 26, 2023

• Merge branch 'ggerganov:master' into master

  tiendung

  authored
  03d1172

• Add logo to README.md

  ggerganov

  authored
  348d692

• Exit from interactive mode if input stream is bad (ggerganov#491)

  haraldF

  authored
  33e35b8

Commits on Mar 25, 2023

• CI: Run other sanitizer builds even if one fails (ggerganov#511)

  anzz1

  authored
  1972616

• Clarify console output in convert-pth-to-ggml.py (ggerganov#512)

  jp-x-g

  authored
  f732695

• CMake / CI additions (ggerganov#497)

  anzz1

  authored
  2f7bf7d

• (Windows) Set console to UTF-8 on init (ggerganov#420)

  anzz1

  authored
  34ab526

• Fix colors enabling on WIN32

  ggerganov

  committed
  c2b25b6

• If n_predict == -1, generate forever

  ggerganov

  committed
  79b2b26

• Inifinite generation via context swapping (ggerganov#71)

  ggerganov

  committed
  e2d490d

• Cleanup STL headers + fix embedding examples + minor stuff

  ggerganov

  committed
  03f7e33

• Move chat scripts into "./examples"

  ggerganov

  committed
  55ad42a

• Add AVX2 implementation of dequantize_row_q4_1 (ggerganov#505)

  slaren

  authored
  459e93c

• Overhaul the examples structure

  ggerganov

  committed
  a316a42

• Retire the ggml_mul_mat() branch for transposed src0 (ggerganov#500)

  ggerganov

  authored
  ecbe466

• Disable prompt verbosity by default and add option to enable (ggerganov#480)

  ggerganov

  committed
  502a400

• Add AVX2 implementation of dequantize_row_q4_0 (ggerganov#467)

  slaren

  authored
  09aecbf

• Don't interefe with BLAS for large prompts by running only 1 thread

  ggerganov

  committed
  4640eff

• Add longer DAN prompt for testing big batch numbers

  ggerganov

  committed
  ab77d76

• Add timings for the prompt evaluation (ggerganov#478)

  slaren

  authored
  29b7baa

• Remove obsolete information from README

  ggerganov

  authored
  4a7129a

• Remove obsolete assert and fix compiler warning

  ggerganov

  committed
  6b6dbc8

• Fix nasty bug in ggml_compute_forward_mul_mat_f32() and reenable BLAS

  ggerganov

  committed
  2a2e63c

• bounds checking for input prefix (ggerganov#492)

  anzz1

  authored
  e899bf5

• feat: '--in-prefix STRING' option (ggerganov#426)

  anzz1

  authored
  fbd4d38

• Add support for file load progress reporting callbacks (ggerganov#434)

  j-f1

  authored
  58e6c9f

• Add missing struct annotation (ggerganov#483)

  Doomsdayrs

  authored
  36d0753

• Fix crash for 65B model with pre-allocated memory (ggerganov#485)

  chriskuehl

  authored
  6f1ee4b

Commits on Mar 24, 2023

• Disable BLAS altogether - the bug is not just for qunatized mat mul

  ggerganov

  committed
  8520fc3

• Disable BLAS branch in mul_mat - seems there is a bug

  ggerganov

  committed
  b3f460e

• Immediately start processing the prompt before user input has been provided (ggerganov#476)

  ggerganov

  authored
  04c6f5e

• Reduce memory usage and allocate enough memory for largest context (ggerganov#473)

  ggerganov

  authored
  7a9b6c3

• Temporary bump the memory buffer size - hopefully fix issues from 483bab2

  ggerganov

  committed
  31572d9

• Update README.md (ggerganov#444)

  gjmulder

  authored
  f4f5362

• fix instruct mode (ggerganov#445)

  rabidcopy

  authored
  863f65e

Pagination

PreviousNext