Commit graph

  • 75418dc2c2 change in spaces, change in fprintf formating Bartosz Podkanowicz 2023-11-09 15:45:51 +01:00
  • db2a5beef1 erase constant seed, add changing alpha and beta parameters from command line Bartosz Podkanowicz 2023-11-09 15:37:53 +01:00
  • 6870c31933 updated docs Concedo 2023-11-09 21:33:19 +08:00
  • cc58ad00b0 Merge branch 'master' into feat-gguf-py-read-refactor KerfuffleV2 2023-11-09 05:25:24 -07:00
  • 6be335678a Update conversion script to convert-hf-to-gguf.py Galunid 2023-11-09 11:45:40 +01:00
  • 2f415527ff Merge branch 'master' into stablelm-support Galunid 2023-11-09 11:44:38 +01:00
  • a75fa576ab
    scripts: Generalize convert scripts (#3838) Galunid 2023-11-09 11:09:29 +01:00
  • 0d0306e7df Include a gguf Python package version bump KerfuffleV2 2023-11-09 02:56:20 -07:00
  • 8e250fe527 Add more information to GGUFReader and examples comments KerfuffleV2 2023-11-09 02:52:42 -07:00
  • c938a1011d updated lite Concedo 2023-11-09 17:21:27 +08:00
  • 7ef4ec3b16 added trim_stop flag Concedo 2023-11-09 16:55:44 +08:00
  • 71c0f3c02f adding in include needed mike dupont 2023-11-09 02:45:02 -05:00
  • 2360aaadb4 Make examples executable, formatting changes KerfuffleV2 2023-11-09 00:25:20 -07:00
  • 855486c912
    Update gguf-py/gguf/gguf_reader.py type hint Kerfuffle 2023-11-09 00:22:00 -07:00
  • 2af29ffeaa
    Update gguf-py/examples/modify_gguf.py formatting Kerfuffle 2023-11-09 00:21:36 -07:00
  • 4a5cd6924f
    Clean up gguf-py/examples/modify_gguf.py whitespace Kerfuffle 2023-11-09 00:21:15 -07:00
  • 557d867278 Minor cleanups. KerfuffleV2 2023-11-09 00:10:14 -07:00
  • afa466807d nooby layer selector considers contextsize Concedo 2023-11-09 14:05:35 +08:00
  • 93e99179be colab updated Concedo 2023-11-09 13:49:06 +08:00
  • 2789ad9392 Fix import path Galunid 2023-11-09 04:37:44 +01:00
  • 89ea740775 Merge branch 'master' into server-completion-edit Jhen 2023-11-09 10:02:10 +08:00
  • 57ad015dc3
    server : add min_p param (#3877) b1500 Mihai 2023-11-09 04:00:34 +02:00
  • af00cca08e Merge commit 'ec893798b7' into HEAD Jared Van Bortel 2023-11-08 16:36:00 -05:00
  • c438c16896 fix build with external fmtlib (v10) Jared Van Bortel 2023-11-06 21:08:48 -05:00
  • a8cac53207 kompute : fix issues with debug layers Jared Van Bortel 2023-11-06 17:24:14 -05:00
  • cba61802c2
    server : fix crash when prompt exceeds context size ZXED 2023-11-08 21:28:39 +03:00
  • faeca572d3
    Adjust versioning in CMakeLists Philipp Zagar 2023-11-08 09:26:05 -08:00
  • 892c6f7069
    Semantic versioning of releases Philipp Zagar 2023-11-08 09:22:34 -08:00
  • 63b3776e1a Fix invalid seqnum in commands when seqnum ommitted in some cases. KerfuffleV2 2023-11-03 15:02:47 -06:00
  • 87061cad81 Remove build-info.h include KerfuffleV2 2023-11-03 05:15:47 -06:00
  • a0c5587401 Expand simple-inference command handling. KerfuffleV2 2023-11-03 05:06:57 -06:00
  • a10f7cd087 Fix logic in simple-inference chunk concat and dump KerfuffleV2 2023-11-02 06:11:52 -06:00
  • e2990ff469 Fix batched-bench directly depending on common.o KerfuffleV2 2023-11-02 04:40:49 -06:00
  • 34175b0b0c Expand simple-inference command support KerfuffleV2 2023-11-02 04:31:39 -06:00
  • 11fa3dfd69 Implementation of a sequence repetition penalty KerfuffleV2 2023-08-12 14:30:45 -06:00
  • b56ed66195 Damagage is not a word. KerfuffleV2 2023-11-08 09:11:22 -07:00
  • fffdac32b5 Fix an issue with state init in GGUFReader KerfuffleV2 2023-11-08 09:01:13 -07:00
  • 0b13ebed6a Merge branch 'master' into concedo_experimental Concedo 2023-11-08 20:54:09 +08:00
  • 73bb25e901
    oai proxy: workaround for some client (such as Chatbox) rhjdvsgsgks 2023-11-08 12:20:40 +00:00
  • 875fb42871
    ggml-alloc : fix backend assignments of views (#3982) b1499 slaren 2023-11-08 13:15:14 +01:00
  • f9deafe825 build: support ppc64le build for make and CMake Roger Meier 2023-11-05 18:32:58 +01:00
  • 38c5b7ee5f fix compilation error for msbuild trt 2023-11-08 03:52:30 +01:00
  • c79c358550 Merge branch 'master' into generic-convert Galunid 2023-11-08 03:49:17 +01:00
  • 2862d16fac Remove outdated scripts Galunid 2023-11-08 03:47:14 +01:00
  • f2292fcc19 fix NamedTuple and Enum usage Jared Van Bortel 2023-11-07 21:12:26 -05:00
  • f364636b2e style cleanup with flake8 Jared Van Bortel 2023-11-07 21:05:41 -05:00
  • ce865b3ce3 Fix missing return statement in add_tensor KerfuffleV2 2023-11-07 18:43:23 -07:00
  • a6f5742a53 sort imports with isort (again) Jared Van Bortel 2023-11-07 20:28:35 -05:00
  • 25d60dcf50 contrastive: initial example Bartosz Podkanowicz 2023-11-08 02:08:45 +01:00
  • d7688dc937 Various type annotation fixes. KerfuffleV2 2023-11-07 17:30:11 -07:00
  • 3469f5a922 ggml-alloc : fix backend assignments of views slaren 2023-11-07 20:31:11 +01:00
  • 6a83bce114 parse_args improvements Galunid 2023-11-07 23:48:17 +01:00
  • b7148838f5 Rename variable Galunid 2023-11-07 23:16:43 +01:00
  • 88b0d9effc Review fixes Galunid 2023-11-07 23:14:58 +01:00
  • 8047aa192f Replay changes from #3871 KerfuffleV2 2023-11-07 15:01:36 -07:00
  • b8c80df741 gguf-py: Refactor and add file reading support KerfuffleV2 2023-11-07 13:56:13 -07:00
  • 4713a40c11 LLAMA_BACKEND_OFFLOAD* -> llama_backend_offload* Galunid 2023-11-07 22:26:31 +01:00
  • 0a7c980b6f
    gguf : track writer state, free unneeded tensors, cleanup (#3871) Jared Van Bortel 2023-11-07 12:43:04 -05:00
  • 413503d4b9
    make : do not add linker flags when compiling static llava lib (#3977) b1497 Georgi Gerganov 2023-11-07 19:25:32 +02:00
  • beb986c942 gguf : bump version to 0.4.6 Jared Van Bortel 2023-11-07 11:25:35 -05:00
  • 548ec463c6
    train : allocate grads for gb_tmp Georgi Gerganov 2023-11-07 16:42:19 +02:00
  • 698c94595e Merge branch 'master' into stablelm-support Galunid 2023-11-07 11:22:13 +01:00
  • 73780f5939 Change ftype from int value to str value Galunid 2023-11-07 10:34:25 +01:00
  • fb3bcac368 handle memory separately for kcpp Concedo 2023-11-07 17:15:14 +08:00
  • a4de8042ee
    train : allocate grads for backward graphs Georgi Gerganov 2023-11-07 10:17:45 +02:00
  • 1b723a8d89
    make : do not add linker flags when compiling static llava lib Georgi Gerganov 2023-11-07 10:13:31 +02:00
  • aa1f36c90a
    Merge branch 'master' into sync Georgi Gerganov 2023-11-07 10:05:19 +02:00
  • e9c1cecb9d
    ggml : fix backward rope after YaRN (#3974) b1496 xaedes 2023-11-07 09:04:51 +01:00
  • 54b4df8886
    Use params when loading models in llava-cli (#3976) b1495 Matthew Tejo 2023-11-06 23:43:59 -08:00
  • 39a80fd368 Use params when loading models in llava-cli matthew 2023-11-06 23:23:39 -08:00
  • f277ed0e8c Merge branch 'master' into concedo_experimental Concedo 2023-11-07 15:23:08 +08:00
  • 46876d2a2c
    cuda : supports running on CPU for GGML_USE_CUBLAS=ON build (#3946) b1494 Meng Zhang 2023-11-06 22:49:08 -08:00
  • 7a3433b4b6 store_true defaults to False, not None Galunid 2023-11-07 07:25:43 +01:00
  • fe7b65ebda server : style improvement Jhen 2023-11-07 10:23:11 +08:00
  • a352d40b5d
    fix function argument type in declaration xaedes 2023-11-07 00:06:34 +01:00
  • ad83e416b3
    add missing function arguments in declaration xaedes 2023-11-07 00:03:57 +01:00
  • 3490a17b15
    fix comments explaining the sinus sign in ggml_forward_rope xaedes 2023-11-06 23:55:09 +01:00
  • 926930ba17
    remove unused ggml_rope_xpos_back xaedes 2023-11-06 23:48:28 +01:00
  • 194e07f62d Remove blank lines Mason M 2023-11-06 18:26:14 -04:00
  • 93e65d32a4
    fix finetune rope call to use correct default attn_factor of 1.0f xaedes 2023-11-06 22:42:39 +01:00
  • ff0a3645d8
    fix backward process of rope xaedes 2023-11-06 22:39:31 +01:00
  • 381efbf480
    llava : expose as a shared library for downstream projects (#3613) b1493 Damian Stewart 2023-11-06 22:36:23 +01:00
  • 354802979e address review comments Jared Van Bortel 2023-11-06 16:07:21 -05:00
  • 05fb6f4e8c sort imports Jared Van Bortel 2023-11-06 15:57:54 -05:00
  • fa234945ed fix defined(...) Meng Zhang 2023-11-06 09:57:27 -08:00
  • 0f3e6b0bba
    fix oai proxy rhjdvsgsgks 2023-11-06 17:55:12 +00:00
  • 03223d6611 Split CPP generation from build-info query Mason M 2023-11-06 13:37:16 -04:00
  • 22f43fca0a Support building as a shared library M. Yusuf Sarıgöz 2023-11-06 19:16:33 +03:00
  • feb60bc447 tokenizer tweaks (+2 squashed commit) Concedo 2023-11-06 21:45:26 +08:00
  • fb8e2fe606 Add support for loading merges.txt KerfuffleV2 2023-10-23 06:27:28 -06:00
  • 372cfef2c3 Merge branch 'concedo' into concedo_experimental Concedo 2023-11-06 20:16:07 +08:00
  • 2102942121 testing LLAMA_PORTABLE flag for building Concedo 2023-11-06 20:15:15 +08:00
  • f36a777bbc convert: Fix handling of LLAMA2 vocab_size = -1 Nikolay Borisov 2023-11-06 13:03:31 +02:00
  • d6d905b242 convert: Fix detection of LLAMA2 Nikolay Borisov 2023-11-06 13:01:00 +02:00
  • 9f03ac7086 Fix according to review comments M. Yusuf Sarıgöz 2023-11-06 13:40:20 +03:00
  • 78ca0667a4 Merge branch 'master' into concedo_experimental Concedo 2023-11-06 16:58:58 +08:00
  • fe381b060b doc: add comments to ggml_cublas_loaded() Meng Zhang 2023-11-05 22:57:43 -08:00
  • 93c4b2a9c6 add force rebuild Concedo 2023-11-06 14:33:42 +08:00
  • 081a86d0df
    Merge branch 'master' into sync Georgi Gerganov 2023-11-06 07:51:40 +02:00
  • 648252ecda Fix flake8 complaints Galunid 2023-11-06 06:28:21 +01:00