Commit Graph

Select branches

Hide Pull Requests

fable5/prefetch-experts

fable5/vram-ondemand

fable5/vram-ondemand-rebase-20260909

fable5/vram-ondemand-rebased-20260726

gemini

Mono Color

• 41318d708e llama : use the same threshold for OpenBLAS and ggml thread limiting (#577) Maël Kerbiriou 2023-03-29 18:10:07 +02:00
• a6956b25a1 add example of re-act pattern (#583) Tobias Lütke 2023-03-29 17:10:24 +02:00
• 83df5639eb Fix GCC warning about binary literal (#595) anzz1 2023-03-29 16:20:07 +03:00
• a5c42c4b13 Fix typo in llama.h (#593) anzz1 2023-03-29 16:19:29 +03:00
• 5a5f8b1501 Enable Fused-Multiply-Add (FMA) and F16C/CVT16 vector extensions on MSVC (#375) anzz1 2023-03-28 22:44:29 +03:00
• f1217055ea CI: fix subdirectory path globbing (#546) anzz1 2023-03-28 22:43:25 +03:00
• 7f4c5c6651 llama : fix linkage with mingw (#551) anzz1 2023-03-28 21:23:09 +03:00
• 2a98bc18ea ggml : add AVX2 implementation of quantize_row_q4_1 (#515) slarenandGeorgi Gerganov 2023-03-28 20:06:03 +02:00
• d0aaff571c py : add temporary script to convert old ggml files to newer version (#539) thementandJakub Horak 2023-03-28 19:55:42 +02:00
• 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 -04:00
• 99c5b27654 ggml : refactor quantized processing functions (#509) Stephan WalterandGeorgi Gerganov 2023-03-28 17:13:01 +00:00
• 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 +09:00
• 96f9c0506f ci : make ctest verbose, hopefully we see what is wrong with the sanitizer Georgi Gerganov 2023-03-28 20:01:09 +03:00
• d502bc7c9d tests : free llama context at the end of the test Georgi Gerganov 2023-03-28 19:51:55 +03:00
• 436e561931 all : be more strict about converting float to double (#458) Stephan WalterandGeorgi Gerganov 2023-03-28 16:48:20 +00:00
• 20e1e84884 deploy : add a Package.swift for SwiftPM support (#393) Jed Fox 2023-03-28 11:39:01 -05:00
• c1f885067c ggml : introduce structs for the q4 data blocks (#356) Stephan WalterandGeorgi Gerganov 2023-03-28 15:56:03 +00:00
• e0670260fb gitignore : add "embedding" Georgi Gerganov 2023-03-28 18:34:35 +03:00
• 28ba975aea Check the existence of f16_model_path_base in quantize.py (#574) dotpy314andJincheng Miao 2023-03-28 23:06:28 +08:00
• a6bdc47cba Fix usage of F16C intrinsics in AVX code (#563) slaren 2023-03-28 16:26:55 +02:00
• 7b8dbcb78b main.cpp fixes, refactoring (#571) anzz1 2023-03-28 17:09:55 +03:00
• 4b8efff0e3 Add embedding example to Makefile (#540) RJ Adriaansen 2023-03-28 08:11:09 +02:00
• 7e5395575a Fix missing ggml link in cmake for examples/* on w64-mingw32 (#542) Marco Matthies 2023-03-27 06:55:26 +02:00
• 34c1072e49 ci: add debug build to sanitizer build matrix (#527) Erik Scholz 2023-03-26 17:48:40 +02:00
• 939ad2d3a5 Fix undefined variables in debug build, remove unused variables (#531) Stephan Walter 2023-03-26 15:34:02 +00:00
• 8c2ec5e21d Add support for linux/arm64 platform during Docker Builds (#514) Juan Calderon-Perez 2023-03-26 10:48:42 -04:00
• b391579db9 Update README and comments for standalone perplexity tool (#525) Stephan Walter 2023-03-26 13:14:01 +00:00
• 7a87d31f4f [main] fix infinite generation (-n == -1) (#523) anzz1 2023-03-26 16:06:10 +03:00
• 348d6926ee Add logo to README.md Georgi Gerganov 2023-03-26 10:20:49 +03:00
• 33e35b8fe8 Exit from interactive mode if input stream is bad (#491) Harald Fernengel 2023-03-26 07:25:46 +02:00
• 19726169b3 CI: Run other sanitizer builds even if one fails (#511) anzz1 2023-03-26 00:13:28 +02:00
• f732695cd5 Clarify console output in convert-pth-to-ggml.py (#512) jp-x-g 2023-03-25 14:53:55 -07:00
• 2f7bf7dd7c CMake / CI additions (#497) anzz1 2023-03-25 23:38:11 +02:00
• 34ab526843 (Windows) Set console to UTF-8 on init (#420) anzz1 2023-03-25 22:29:22 +02:00
• c2b25b6912 Fix colors enabling on WIN32 Georgi Gerganov 2023-03-25 21:53:39 +02:00
• 79b2b266db If n_predict == -1, generate forever Georgi Gerganov 2023-03-25 21:51:41 +02:00
• e2d490dafd Inifinite generation via context swapping (#71) Georgi Gerganov 2023-03-25 21:36:22 +02:00
• 03f7e33560 Cleanup STL headers + fix embedding examples + minor stuff Georgi Gerganov 2023-03-25 20:51:14 +02:00
• 55ad42af84 Move chat scripts into "./examples" Georgi Gerganov 2023-03-25 20:36:52 +02:00
• 459e93cce0 Add AVX2 implementation of dequantize_row_q4_1 (#505) slaren 2023-03-25 19:31:48 +01:00
• a316a425d0 Overhaul the examples structure Georgi Gerganov 2023-03-25 20:26:40 +02:00
• ecbe466a36 Retire the ggml_mul_mat() branch for transposed src0 (#500) Georgi Gerganov 2023-03-25 19:47:21 +02:00
• 502a400192 Disable prompt verbosity by default and add option to enable (#480) Georgi Gerganov 2023-03-25 17:16:50 +02:00
• 09aecbf628 Add AVX2 implementation of dequantize_row_q4_0 (#467) slaren 2023-03-25 16:06:49 +01:00
• 4640eff23d Don't interefe with BLAS for large prompts by running only 1 thread Georgi Gerganov 2023-03-25 17:03:10 +02:00
• ab77d76312 Add longer DAN prompt for testing big batch numbers Georgi Gerganov 2023-03-25 16:47:59 +02:00
• 29b7baab67 Add timings for the prompt evaluation (#478) slaren 2023-03-25 15:34:23 +01:00
• 4a7129acd2 Remove obsolete information from README Georgi Gerganov 2023-03-25 16:30:32 +02:00
• 6b6dbc8910 Remove obsolete assert and fix compiler warning Georgi Gerganov 2023-03-25 16:22:05 +02:00
• 2a2e63ce05 Fix nasty bug in ggml_compute_forward_mul_mat_f32() and reenable BLAS Georgi Gerganov 2023-03-25 16:09:54 +02:00
• e899bf54b2 bounds checking for input prefix (#492) anzz1 2023-03-25 14:42:09 +02:00
• fbd4d38c64 feat: '--in-prefix STRING' option (#426) anzz1 2023-03-25 14:03:19 +02:00
• 58e6c9f36f Add support for file load progress reporting callbacks (#434) Jed Fox 2023-03-25 01:26:28 -04:00
• 36d07532ef Add missing struct annotation (#483) Doomsdayrs 2023-03-25 01:21:24 -04:00
• 6f1ee4b640 Fix crash for 65B model with pre-allocated memory (#485) Chris Kuehl 2023-03-24 23:38:14 -05:00
• 8520fc310e Disable BLAS altogether - the bug is not just for qunatized mat mul Georgi Gerganov 2023-03-24 23:47:06 +02:00
• b3f460e941 Disable BLAS branch in mul_mat - seems there is a bug Georgi Gerganov 2023-03-24 23:39:17 +02:00
• 04c6f5ed6f Immediately start processing the prompt before user input has been provided (#476) Georgi Gerganov 2023-03-24 23:17:58 +02:00
• 7a9b6c3a8b Reduce memory usage and allocate enough memory for largest context (#473) Georgi Gerganov 2023-03-24 23:17:37 +02:00
• 31572d9665 Temporary bump the memory buffer size - hopefully fix issues from 483bab2e Georgi Gerganov 2023-03-24 18:23:56 +02:00
• f4f5362edb Update README.md (#444) Gary Mulder 2023-03-24 15:23:09 +00:00
• 863f65e2e3 fix instruct mode (#445) rabidcopy 2023-03-24 10:22:39 -05:00
• afd220d9c6 Properly free llama_context on failure Georgi Gerganov 2023-03-24 17:21:01 +02:00
• 481044d50c additional optimizations for POWER9 (#454) Cameron Kaiser 2023-03-24 08:19:26 -07:00
• 563cdc391d Support calling mlock() on loaded model data on Linux and macOS (#453) comexandGeorgi Gerganov 2023-03-24 08:19:05 -07:00
• 8d4a855c24 Add embedding mode with arg flag. Currently working (#282) LucianoandGeorgi Gerganov 2023-03-24 08:05:13 -07:00
• b6b268d441 Add link to Roadmap discussion Georgi Gerganov 2023-03-24 09:13:35 +02:00
• 3cd8dde0d1 Revert "Fix memory allocation issues and seg faults" Georgi Gerganov 2023-03-24 06:22:28 +02:00
• 4870e455b3 Fix memory allocation issues and seg faults Georgi Gerganov 2023-03-24 00:11:53 +02:00
• 483bab2e3d Avoid the transposed X branch in the Z = X * Y matrix multiplication (#439) Georgi Gerganov 2023-03-23 23:22:01 +02:00
• 404e1da38e Fix quantize script not finding models in parent directory (#428) Jed Fox 2023-03-23 16:42:52 -04:00
• 4cc053b6d5 Remove oboslete command from Docker script Georgi Gerganov 2023-03-23 22:39:44 +02:00
• 0ba5a3a9a5 Obsolete Georgi Gerganov 2023-03-23 22:32:02 +02:00
• 2e17dfd80a Replace EOS with newline to prevent context/memory being flushed by EOS in interactive mode (#333) 3 people 2023-03-23 15:22:47 -05:00
• 20a1a4e09c Fix GPTQ converter (#423) Timmy KnightandGeorgi Gerganov 2023-03-23 10:18:13 -10:00
• ad072fc5ad Generate library with CMake (#430) nusu-github 2023-03-24 05:16:48 +09:00
• ea10d3ded2 Command line args bounds checking (#424) anzz1 2023-03-23 19:54:28 +02:00
• a18c19259a Fix Nix build Ben Siraphob 2023-03-22 00:37:02 -05:00
• a50e39c6fe Revert "Delete SHA256SUMS for now" (#429) Stephan WalterandPavol Rusnak 2023-03-23 14:15:48 +00:00
• a140219e81 Fix Makefile echo escape codes (by removing them). (#418) Kerfuffle 2023-03-23 05:41:32 -06:00
• 8a3e5ef801 Move model section from issue template to README.md (#421) Gary Mulder 2023-03-23 11:30:40 +00:00
• 8eea5ae0e5 Delete SHA256SUMS for now (#416) anzz1 2023-03-23 12:26:19 +02:00
• 93208cfb92 Adjust repetition penalty .. Georgi Gerganov 2023-03-23 10:46:58 +02:00
• 03ace14cfd Add link to recent podcast about whisper.cpp and llama.cpp Georgi Gerganov 2023-03-23 09:48:51 +02:00
• e4412b45e3 CI: CMake: Separate build and test steps (#376) anzz1 2023-03-23 04:20:34 +02:00
• f7dc43bc0d Fix instruct mode broken by PR #354 (#409) tjohnmanandJohnman 2023-03-23 01:30:23 +01:00
• ee8a788786 Update issue template so people will use it (#404) Gary Mulder 2023-03-22 19:06:18 +00:00
• 69c92298a9 Deduplicate q4 quantization functions (#383) Stephan WalterandGeorgi Gerganov 2023-03-22 17:29:06 +00:00
• 97940520e8 fix: add POSIX functionality for Linux compilation (#51) Valentyn Bezshapkin 2023-03-22 18:20:25 +01:00
• 305ba6f0e6 Don't force immediate interactive without -i (#354) tjohnmanandJohnman 2023-03-22 18:16:35 +01:00
• 4122dffff9 cmake: make llama an actual library (#392) Erik Scholz 2023-03-22 17:37:10 +01:00
• 56e659a0b2 fix perplexity after c-api refactor (#390) Erik Scholz 2023-03-22 17:09:38 +01:00
• 40ea807a97 Add details on perplexity to README.md (#395) Gary Linscott 2023-03-22 08:53:54 -07:00
• d5850c53ca Add missing header for memcpy (#386) Yusuf Kağan Hanoğlu 2023-03-22 11:55:45 +03:00
• ae44e23ee3 When seed <= 0 - use the clock to generate one Georgi Gerganov 2023-03-22 07:47:15 +02:00
• 928480ef5b Init llama_context_params properly from CLI (#370) Georgi Gerganov 2023-03-22 07:45:00 +02:00
• 56817b1f88 Remove temporary notice and update hot topics Georgi Gerganov 2023-03-22 07:34:02 +02:00
• f5a77a629b Introduce C-style API (#370) Georgi Gerganov 2023-03-22 07:32:36 +02:00
• da0e9fe90c Add SHA256SUMS file and instructions to README how to obtain and verify the downloads Gary Mulder 2023-03-20 20:14:06 +00:00
• e6c9e0986c Fix bin dir for win ci anzz1 2023-03-21 23:49:24 +02:00