Commit graph

Select branches

Hide pull requests

master

Mono Color

• e986f94829
  Added api for getting/setting the kv_cache (#685) Christian Falch 2023-04-02 12:23:04 +0200
• c0bb1d3ce2
  ggml : change ne to int64_t (#626) Marian Cepok 2023-04-02 12:21:31 +0200
• 6e7801d08d
  examples : add gpt4all script (#658) Leonardo Neumann 2023-04-02 04:56:20 -0300
• 81040f10aa
  llama : do not allocate KV cache for "vocab_only == true" (#682) Stephan Walter 2023-04-02 07:18:53 +0000
• c4f89d8d73
  make : use -march=native -mtune=native on x86 (#609) Fabian 2023-04-02 09:17:05 +0200
• 5b70e7de4c
  fix default params for examples/main (#697) Murilo Santana 2023-04-01 23:41:12 -0300
• a717cba844
  py: huggingface -> Hugging Face (#686) Ikko Eltociear Ashimine 2023-04-02 01:38:18 +0900
• d0a7f742e7
  readme: replace termux links with homepage, play store is deprecated (#680) rimoliga 2023-04-01 11:57:30 -0300
• 0d054e292e Show error message when -f fails Slaren 2023-03-31 20:03:48 +0200
• 3525899277
  Enable -std= for cmake builds, fix warnings (#598) Stephan Walter 2023-03-31 19:19:16 +0000
• 1d08882afa
  Optimize AVX2 ggml_vec_dot_q4_0 (#642) slaren 2023-03-31 17:55:52 +0200
• 02c5b27e91
  Add AVX acceleration (#617) perserk 2023-03-31 16:55:44 +0500
• cbef542879 py : cleanup the code Pavol Rusnak 2023-03-29 21:31:24 +0200
• 9733104be5 drop quantize.py (now that models are using a single file) Pavol Rusnak 2023-03-31 00:52:06 +0200
• 3df890aef4
  readme : update supported models Georgi Gerganov 2023-03-30 22:31:54 +0300
• ee0c40dd6d Introduce GGML migration tool for new file format Justine Tunney 2023-03-30 05:42:56 -0700
• 6f23ba5ee2 Ensure --mlock works properly with mmap() support Justine Tunney 2023-03-30 01:53:36 -0700
• 78ca9838ee Make loading weights 10-100x faster Justine Tunney 2023-03-29 13:51:37 -0700
• a017390358 Initial windows support (untested) Slaren 2023-03-29 22:22:36 +0200
• ac184d5147 Always initialize mm_addr and mm_length in llama_model Slaren 2023-03-29 08:53:14 +0200
• 276e5b7811 Unmap the file in llama_free Slaren 2023-03-29 08:31:26 +0200
• d68c5dc435 Make mmap_file static Slaren 2023-03-29 06:18:18 +0200
• 64bde3ffd4 Fix ggml_init_params in quantize Slaren 2023-03-29 05:38:57 +0200
• c03ae8dca1 Add mmap support for model files Slaren 2023-03-29 02:03:43 +0200
• 3bcc129ba8
  cmake : properly invoke CTest (#629) Stephan Walter 2023-03-30 17:56:59 +0000
• a4755cf288
  Remove unused variable (#607) Casey Primozic 2023-03-30 10:53:35 -0700
• 1f0414feec
  make : fix darwin f16c flags check (#615) david raistrick 2023-03-30 13:34:45 -0400
• 77efdf5a50
  ggml : fix NEON signs (close #620, #622) Georgi Gerganov 2023-03-30 20:27:32 +0300
• ed3c680bcd
  Fix GGML_F32Cx8_STORE in AVX without F16C path (#619) slaren 2023-03-30 11:16:30 +0200
• 9cbc404ba6
  ci : re-enable AVX512 testing (Windows-MSVC) (#584) anzz1 2023-03-29 23:44:39 +0300
• b51c717d5c
  ggml : init time on first ggml_init() call Georgi Gerganov 2023-03-29 22:15:34 +0300
• 0ba76c1e73
  llama : fix compile warnings when reading the vocab Georgi Gerganov 2023-03-29 22:13:12 +0300
• cea1c85948
  ggml : add ARM_NEON dequantize_row_q4_1() Georgi Gerganov 2023-03-29 22:10:01 +0300
• f202ada131
  ggml : add ARM_NEON quantize_row_q4_1() Georgi Gerganov 2023-03-29 22:03:02 +0300
• 3b44d30d9b
  ggml : add ARM_NEON ggml_vec_dot_q4_1() Georgi Gerganov 2023-03-29 21:47:33 +0300
• 61cbfff5c9
  rename convert_ggml_to_pth.py -> convert-ggml-to-pth.py (#600) Pavol Rusnak 2023-03-29 20:09:25 +0200
• d9ad104440
  Create chat-13B.bat (#592) Thérence 2023-03-29 19:21:09 +0200
• b467702b87
  readme : fix typos Georgi Gerganov 2023-03-29 19:38:31 +0300
• 516d88e75c
  readme : add GPT4All instructions (close #588) Georgi Gerganov 2023-03-29 19:37:20 +0300
• 53635c081c
  py : add GPT4All conversion script Georgi Gerganov 2023-03-29 19:29:26 +0300
• 41318d708e
  llama : use the same threshold for OpenBLAS and ggml thread limiting (#577) Maël Kerbiriou 2023-03-29 18:10:07 +0200
• a6956b25a1
  add example of re-act pattern (#583) Tobias Lütke 2023-03-29 17:10:24 +0200
• 83df5639eb
  Fix GCC warning about binary literal (#595) anzz1 2023-03-29 16:20:07 +0300
• a5c42c4b13
  Fix typo in llama.h (#593) anzz1 2023-03-29 16:19:29 +0300
• 5a5f8b1501
  Enable Fused-Multiply-Add (FMA) and F16C/CVT16 vector extensions on MSVC (#375) anzz1 2023-03-28 22:44:29 +0300
• f1217055ea
  CI: fix subdirectory path globbing (#546) anzz1 2023-03-28 22:43:25 +0300
• 7f4c5c6651
  llama : fix linkage with mingw (#551) anzz1 2023-03-28 21:23:09 +0300
• 2a98bc18ea
  ggml : add AVX2 implementation of quantize_row_q4_1 (#515) slaren 2023-03-28 20:06:03 +0200
• d0aaff571c
  py : add temporary script to convert old ggml files to newer version (#539) thement 2023-03-28 19:55:42 +0200
• d0330fd783
  py : add capabiliy to convert from ggml back to torch or hf format for further consumption/training/finetuning (#403) Tai Duc Nguyen 2023-03-28 13:51:29 -0400
• 99c5b27654
  ggml : refactor quantized processing functions (#509) Stephan Walter 2023-03-28 17:13:01 +0000
• 692ce3164e
  py : removed unused model variable and verified that the code functions correctly with vocab_only setting. Also confirmed that the code works as expected after running with reduced memory usage due to deletion of no-longer-needed variable. (#547) DooWoong Lee (David) 2023-03-29 02:02:34 +0900
• 96f9c0506f
  ci : make ctest verbose, hopefully we see what is wrong with the sanitizer Georgi Gerganov 2023-03-28 20:01:09 +0300
• d502bc7c9d
  tests : free llama context at the end of the test Georgi Gerganov 2023-03-28 19:51:55 +0300
• 436e561931
  all : be more strict about converting float to double (#458) Stephan Walter 2023-03-28 16:48:20 +0000
• 20e1e84884
  deploy : add a Package.swift for SwiftPM support (#393) Jed Fox 2023-03-28 11:39:01 -0500
• c1f885067c
  ggml : introduce structs for the q4 data blocks (#356) Stephan Walter 2023-03-28 15:56:03 +0000
• e0670260fb
  gitignore : add "embedding" Georgi Gerganov 2023-03-28 18:34:35 +0300
• 28ba975aea
  Check the existence of f16_model_path_base in quantize.py (#574) dotpy314 2023-03-28 23:06:28 +0800
• a6bdc47cba
  Fix usage of F16C intrinsics in AVX code (#563) slaren 2023-03-28 16:26:55 +0200
• 7b8dbcb78b
  main.cpp fixes, refactoring (#571) anzz1 2023-03-28 17:09:55 +0300
• 4b8efff0e3
  Add embedding example to Makefile (#540) RJ Adriaansen 2023-03-28 08:11:09 +0200
• 7e5395575a
  Fix missing ggml link in cmake for examples/* on w64-mingw32 (#542) Marco Matthies 2023-03-27 06:55:26 +0200
• 34c1072e49
  ci: add debug build to sanitizer build matrix (#527) Erik Scholz 2023-03-26 17:48:40 +0200
• 939ad2d3a5
  Fix undefined variables in debug build, remove unused variables (#531) Stephan Walter 2023-03-26 15:34:02 +0000
• 8c2ec5e21d
  Add support for linux/arm64 platform during Docker Builds (#514) Juan Calderon-Perez 2023-03-26 10:48:42 -0400
• b391579db9
  Update README and comments for standalone perplexity tool (#525) Stephan Walter 2023-03-26 13:14:01 +0000
• 7a87d31f4f
  [main] fix infinite generation (-n == -1) (#523) anzz1 2023-03-26 16:06:10 +0300
• 348d6926ee
  Add logo to README.md Georgi Gerganov 2023-03-26 10:20:49 +0300
• 33e35b8fe8
  Exit from interactive mode if input stream is bad (#491) Harald Fernengel 2023-03-26 07:25:46 +0200
• 19726169b3
  CI: Run other sanitizer builds even if one fails (#511) anzz1 2023-03-26 00:13:28 +0200
• f732695cd5
  Clarify console output in convert-pth-to-ggml.py (#512) jp-x-g 2023-03-25 14:53:55 -0700
• 2f7bf7dd7c
  CMake / CI additions (#497) anzz1 2023-03-25 23:38:11 +0200
• 34ab526843
  (Windows) Set console to UTF-8 on init (#420) anzz1 2023-03-25 22:29:22 +0200
• c2b25b6912
  Fix colors enabling on WIN32 Georgi Gerganov 2023-03-25 21:53:39 +0200
• 79b2b266db
  If n_predict == -1, generate forever Georgi Gerganov 2023-03-25 21:51:41 +0200
• e2d490dafd
  Inifinite generation via context swapping (#71) Georgi Gerganov 2023-03-25 21:36:22 +0200
• 03f7e33560
  Cleanup STL headers + fix embedding examples + minor stuff Georgi Gerganov 2023-03-25 20:51:14 +0200
• 55ad42af84
  Move chat scripts into "./examples" Georgi Gerganov 2023-03-25 20:36:52 +0200
• 459e93cce0
  Add AVX2 implementation of dequantize_row_q4_1 (#505) slaren 2023-03-25 19:31:48 +0100
• a316a425d0
  Overhaul the examples structure Georgi Gerganov 2023-03-25 20:26:40 +0200
• ecbe466a36
  Retire the ggml_mul_mat() branch for transposed src0 (#500) Georgi Gerganov 2023-03-25 19:47:21 +0200
• 502a400192
  Disable prompt verbosity by default and add option to enable (#480) Georgi Gerganov 2023-03-25 17:16:50 +0200
• 09aecbf628
  Add AVX2 implementation of dequantize_row_q4_0 (#467) slaren 2023-03-25 16:06:49 +0100
• 4640eff23d
  Don't interefe with BLAS for large prompts by running only 1 thread Georgi Gerganov 2023-03-25 17:03:10 +0200
• ab77d76312
  Add longer DAN prompt for testing big batch numbers Georgi Gerganov 2023-03-25 16:47:59 +0200
• 29b7baab67
  Add timings for the prompt evaluation (#478) slaren 2023-03-25 15:34:23 +0100
• 4a7129acd2
  Remove obsolete information from README Georgi Gerganov 2023-03-25 16:30:32 +0200
• 6b6dbc8910
  Remove obsolete assert and fix compiler warning Georgi Gerganov 2023-03-25 16:22:05 +0200
• 2a2e63ce05
  Fix nasty bug in ggml_compute_forward_mul_mat_f32() and reenable BLAS Georgi Gerganov 2023-03-25 16:09:54 +0200
• e899bf54b2
  bounds checking for input prefix (#492) anzz1 2023-03-25 14:42:09 +0200
• fbd4d38c64
  feat: '--in-prefix STRING' option (#426) anzz1 2023-03-25 14:03:19 +0200
• 58e6c9f36f
  Add support for file load progress reporting callbacks (#434) Jed Fox 2023-03-25 01:26:28 -0400
• 36d07532ef
  Add missing struct annotation (#483) Doomsdayrs 2023-03-25 01:21:24 -0400
• 6f1ee4b640
  Fix crash for 65B model with pre-allocated memory (#485) Chris Kuehl 2023-03-24 23:38:14 -0500
• 8520fc310e
  Disable BLAS altogether - the bug is not just for qunatized mat mul Georgi Gerganov 2023-03-24 23:47:06 +0200
• b3f460e941
  Disable BLAS branch in mul_mat - seems there is a bug Georgi Gerganov 2023-03-24 23:39:17 +0200
• 04c6f5ed6f
  Immediately start processing the prompt before user input has been provided (#476) Georgi Gerganov 2023-03-24 23:17:58 +0200
• 7a9b6c3a8b
  Reduce memory usage and allocate enough memory for largest context (#473) Georgi Gerganov 2023-03-24 23:17:37 +0200
• 31572d9665
  Temporary bump the memory buffer size - hopefully fix issues from 483bab2e Georgi Gerganov 2023-03-24 18:23:56 +0200