Commit Graph

  • 3e2327c96a
    Try manually adding CXX flag beiller 2023-03-08 16:44:50 -0500
  • 3c04dfb436
    run build in shell beiller 2023-03-08 16:44:50 -0500
  • 7deae8a2ca
    fix build procedure beiller 2023-03-08 16:44:50 -0500
  • 67b1c842d9
    Use sentencepiece tokenization beiller 2023-03-08 16:44:50 -0500
  • 96dc6a0c68
    work towards tokenizer integration beiller 2023-03-08 16:44:50 -0500
  • 0ac8651bd6
    10% performance boost on ARM Georgi Gerganov 2023-03-13 01:30:53 +0200
  • c80e2a8f2a
    Revert "10% performance boost on ARM" Georgi Gerganov 2023-03-13 01:28:08 +0200
  • 54a0e66ea0
    Check for vdotq_s32 availability Georgi Gerganov 2023-03-13 01:21:03 +0200
  • 543c57e991
    Ammend to previous commit - forgot to update non-QRDMX branch Georgi Gerganov 2023-03-13 01:05:24 +0200
  • 113a9e83eb
    10% performance boost on ARM Georgi Gerganov 2023-03-13 00:56:10 +0200
  • 404fac0d62
    Fix color getting reset before prompt output done (#65) Matvey Soloviev 2023-03-12 23:07:34 +0100
  • ea840344b3 Fix color getting reset before prompt output done Matvey Soloviev 2023-03-12 22:50:21 +0100
  • 1a0a74300f
    Update README.md Georgi Gerganov 2023-03-12 23:39:01 +0200
  • 96ea727f47
    Add interactive mode (#61) Matvey Soloviev 2023-03-12 22:13:28 +0100
  • 0453507450 Fix OS X build Matvey Soloviev 2023-03-12 21:59:32 +0100
  • f47face9a5
    Update README to explain interactive mode. Matvey Soloviev 2023-03-12 21:56:45 +0100
  • 798428465c Improve interactive mode. Make rev. prompt optional. Matvey Soloviev 2023-03-12 21:09:04 +0100
  • 81d09b2101 Initial work on interactive mode. Matvey Soloviev 2023-03-12 20:20:45 +0100
  • 9661954835
    Fix typo in README (#45) Marc Köhlbrugge 2023-03-13 03:30:08 +0700
  • f385f8dee8
    Allow using prompt files (#59) Ben Garney 2023-03-12 13:28:36 -0700
  • 63a8b00d88 fix: older standard for compatibility valentynbez 2023-03-12 21:27:22 +0100
  • 02f0c6fe7f
    Add back top_k (#56) beiller 2023-03-12 16:23:15 -0400
  • 90b9b5b8c0
    Update utils.h Georgi Gerganov 2023-03-12 22:22:56 +0200
  • 2913d75a42
    Update utils.cpp Georgi Gerganov 2023-03-12 22:22:37 +0200
  • eb062bb012
    Windows fixes (#31) Sebastián A 2023-03-12 17:15:00 -0300
  • 7027a97837
    Update README.md Georgi Gerganov 2023-03-12 22:09:26 +0200
  • 2d555e5b42
    Add CI (#60) Georgi Gerganov 2023-03-12 22:08:24 +0200
  • 2a3d4c8c7e
    Add CI Georgi Gerganov 2023-03-12 22:06:19 +0200
  • 25c1d0a291
    Update utils.cpp Ben Garney 2023-03-12 12:58:05 -0700
  • 7c9e54e55e
    Revert "weights_only" arg - this causing more trouble than help Georgi Gerganov 2023-03-12 20:59:01 +0200
  • 7412189fe5
    Add back top_k Bill Hamilton 2023-03-08 16:44:50 -0500
  • 3b20e78c16 Change the use of vector for stack allocations. Sebastián Aedo 2023-03-12 12:39:57 -0300
  • 526ddc0101 fix: add POSIX functionality for Linux compilation valentynbez 2023-03-12 15:42:25 +0100
  • b577e34c1d
    Add osx-specific python install instructions John Albietz 2023-03-12 09:00:34 -0400
  • b9bd1d0141
    python/pytorch compat notes (#44) Oleksandr Nikitin 2023-03-12 14:16:33 +0200
  • 2fed42d9d4
    Fix typo in README Marc Köhlbrugge 2023-03-12 18:10:24 +0700
  • 97a25c1460
    python/pytorch compat notes Oleksandr Nikitin 2023-03-12 12:52:41 +0200
  • 640fd77a25 Use buffering maekawatoshiki 2023-03-12 19:38:25 +0900
  • 129c7d1ea8
    Add repetition penalty (#20) beiller 2023-03-12 05:27:42 -0400
  • ebb357f711
    Minor formatting Georgi Gerganov 2023-03-12 11:26:48 +0200
  • 702fddf5c5
    Clarify meaning of hacking Georgi Gerganov 2023-03-12 09:03:25 +0200
  • c2201a9a83 Add comment. Jay Krell 2023-03-11 22:55:17 -0800
  • bea4af5c9c Fix sizeof(pp). Jay Krell 2023-03-11 22:45:08 -0800
  • 7d86e25bf6
    README: add "Supported platforms" + update hot topics Georgi Gerganov 2023-03-12 08:41:54 +0200
  • a93120236f
    use weights_only in conversion script (#32) deepdiffuser 2023-03-11 22:36:35 -0800
  • 3225d9b8a0 nologo Jay Krell 2023-03-11 22:36:05 -0800
  • 6a9a67f0be
    Add LICENSE (#21) Pavol Rusnak 2023-03-12 07:36:03 +0100
  • ec64cfa1db define NDEBUG, and simplify casts. Jay Krell 2023-03-11 22:34:06 -0800
  • a05225f194 LTCG and Win32 is Win32, not specific compilers. Jay Krell 2023-03-11 22:27:38 -0800
  • 636d56818a Port to Visual C++. Jay Krell 2023-03-11 22:00:20 -0800
  • c505a2d217 Fix zero initialization of the other fields. Sebastián Aedo 2023-03-12 02:12:53 -0300
  • 1ed5c7c75b use weights_only in conversion script deepdiffuser 2023-03-11 20:09:28 -0800
  • ce0d5fb9de MSVC: Remove features that are only available on MSVC C++20. Sebastián Aedo 2023-03-12 00:11:08 -0300
  • 718bb989a5 Remove unsupported VLAs Sebastián Aedo 2023-03-12 00:02:03 -0300
  • 2d29d4b89e Apply fixes suggested to build on windows Sebastián Aedo 2023-03-11 19:26:20 -0300
  • 340bff0f0e
    Update comments, more proper application beiller 2023-03-11 21:51:03 -0500
  • da1a4ff01f
    Update README.md Georgi Gerganov 2023-03-12 01:26:32 +0200
  • 4726e671e6
    Remove Unprintable beiller 2023-03-11 17:13:28 -0500
  • 5e625ea07a
    More robust unprintable character check beiller 2023-03-11 17:06:39 -0500
  • e236dbb1e9
    Remove unprintable characters from vocab list beiller 2023-03-11 16:32:21 -0500
  • fc3f02e3c3
    Add LICENSE Pavol Rusnak 2023-03-11 21:31:56 +0100
  • c90e78edc3
    Numeric fix beiller 2023-03-11 14:55:57 -0500
  • 78651d5792
    Update utils.cpp beiller 2023-03-11 14:24:32 -0500
  • 3f6a118d6a
    Update utils.h beiller 2023-03-11 14:24:12 -0500
  • d2b96b5593
    Adding repeat penalization beiller 2023-03-11 14:23:33 -0500
  • faad7f1464
    Add oneliner for batch quantization Juraj Bednar 2023-03-11 18:44:59 +0100
  • 6b2cb6302f
    Fix a typo in model name (#16) Juraj Bednar 2023-03-11 18:32:20 +0100
  • d3bbccc1d7
    Fix a typo in model name Juraj Bednar 2023-03-11 18:28:23 +0100
  • 4235e3d5b3
    Update README.md Georgi Gerganov 2023-03-11 18:10:18 +0200
  • f1eaff4721 Add AVX2 support for x86 architectures thanks to @Const-me ! Georgi Gerganov 2023-03-11 17:58:18 +0200
  • a9e58529ea Fix un-initialized FP16 tables on x86 (#15, #2) Georgi Gerganov 2023-03-11 17:40:14 +0200
  • 7d9ed7b25f
    Bump memory buffer Georgi Gerganov 2023-03-11 12:44:21 +0200
  • 0c6803321c
    Update README.md Georgi Gerganov 2023-03-11 12:31:21 +0200
  • f60fa9e50a
    .gitignore models/ Georgi Gerganov 2023-03-11 12:26:46 +0200
  • 7211862c94
    Update Makefile var + add comment Georgi Gerganov 2023-03-11 12:26:16 +0200
  • a5c5ae2f54
    Update README.md Georgi Gerganov 2023-03-11 11:34:25 +0200
  • ea977e85ec
    Update README.md Georgi Gerganov 2023-03-11 11:34:11 +0200
  • 007a8f6f45
    Support all LLaMA models + change Q4_0 quantization storage Georgi Gerganov 2023-03-11 10:47:09 +0200
  • 5f2f970d51
    Include Python dependencies in README (#6) Simon Willison 2023-03-10 21:47:26 -0800
  • 029f2b1fcd
    Include Python dependencies in README Simon Willison 2023-03-10 20:26:36 -0800
  • 73c6ed5e87
    Update README.md Georgi Gerganov 2023-03-11 01:30:47 +0200
  • 01eeed8fb1
    Update README.md Georgi Gerganov 2023-03-11 01:22:58 +0200
  • 6da2df34ee
    Update README.md Georgi Gerganov 2023-03-11 01:18:10 +0200
  • 9dcf4dba45
    Add missing headers for memcpy and assert (#3) Jean-Michaël Celerier 2023-03-10 18:04:06 -0500
  • 920a7fe2d9
    Update README.md Georgi Gerganov 2023-03-11 00:55:22 +0200
  • df8abc8d29
    Add missing headers for memcpy and assert Jean-Michaël Celerier 2023-03-10 17:51:54 -0500
  • 3a57ee59de
    Update README.md Georgi Gerganov 2023-03-11 00:51:46 +0200
  • b85028522d
    Update README.md Georgi Gerganov 2023-03-11 00:09:19 +0200
  • 8a01f565ff
    Update README.md Georgi Gerganov 2023-03-10 23:53:11 +0200
  • 70bc0b8b15
    Fix a bug in the rope calculation Georgi Gerganov 2023-03-10 23:46:39 +0200
  • 18ebda34d6
    Update README.md Georgi Gerganov 2023-03-10 21:52:27 +0200
  • 319cdb3e1f
    Final touches Georgi Gerganov 2023-03-10 21:50:46 +0200
  • 775328064e
    Create README.md Georgi Gerganov 2023-03-10 21:47:46 +0200
  • 26c0846629
    Initial release Georgi Gerganov 2023-03-10 20:40:58 +0200