Commit Graph

  • 85399c679b
    Remove error causing lines for cuBLAS on Windows RelatedTitle 2023-05-02 11:19:24 -0600
  • a9dcc936de
    Add windows build with cuBLAS RelatedTitle 2023-05-02 11:17:36 -0600
  • 0bf680fea2
    talk-llama : fix session prompt load (#854) Luis Herrera 2023-05-02 12:05:27 -0500
  • b806420873
    whisper : add detect-language mode (#853) CRD716 2023-05-02 11:51:52 -0500
  • 3c7d04d0d3 Merge remote-tracking branch 'origin/master' DGdev91 2023-05-02 02:50:47 +0200
  • 66f1c2fdaf Update elevenlabs example to use official python API DGdev91 2023-04-30 01:00:27 +0200
  • 7d0f575604 talk-llama: fixing session prompts load herrera-luis 2023-05-01 19:03:44 -0500
  • edb1295a55
    fix language being english on print CRD716 2023-05-01 18:27:34 -0500
  • b0a567f671
    fix quitting CRD716 2023-05-01 18:23:56 -0500
  • 57ad2c9758
    various fixes CRD716 2023-05-01 18:00:42 -0500
  • 59f5c04b48
    help message fix CRD716 2023-05-01 17:58:02 -0500
  • 62e35efa54
    run language detection if dl is set CRD716 2023-05-01 17:56:27 -0500
  • 5bad3366b5
    no idea why that last one didn't commit CRD716 2023-05-01 17:54:02 -0500
  • 706f2fd5e1
    renaming and help CRD716 2023-05-01 17:53:31 -0500
  • e49189e9bb
    add detectlanguage flag CRD716 2023-05-01 17:49:23 -0500
  • be5911a9f3
    talk-llama : add --session support (#845) Luis Herrera 2023-05-01 12:18:10 -0500
  • cb51ca9b21 talk-llama: adding missing function which updates session_tokens herrera-luis 2023-05-01 11:34:50 -0500
  • e524df5294 talk-llama: update the value of need_to_save_session to true in order to save the session in the subsequent interaction herrera-luis 2023-05-01 11:15:14 -0500
  • e9bf002a40 readme: updating session doc herrera-luis 2023-05-01 09:27:21 -0500
  • 7b9a18ffa1 llama: adding session fixes herrera-luis 2023-05-01 09:03:20 -0500
  • d375d73b2e
    bench : improve benchmarks Georgi Gerganov 2023-05-01 14:44:39 +0300
  • 7765770f89
    whisper : add memory sizes for Q8_0 (close #846) Georgi Gerganov 2023-05-01 10:03:56 +0300
  • 872a85ae94
    whisper.wasm : fix typo in readme (#832) Baffin Lee 2023-05-01 14:28:05 +0800
  • 944ef1192c readme: adding --session info in examples/talk-llama herrera-luis 2023-04-30 21:52:45 -0500
  • fe1ea36d2c feat: adding session support herrera-luis 2023-04-30 21:42:10 -0500
  • 9c61f5f585
    release : v1.4.1 v1.4.1 Georgi Gerganov 2023-04-30 22:57:42 +0300
  • c94c469592
    whisper : fix quantize bug (#842) Georgi Gerganov 2023-04-30 22:50:04 +0300
  • 4e94a20401
    whisper : fix bug during quantization Georgi Gerganov 2023-04-30 22:48:49 +0300
  • 327ab5c1a2
    whisper : debug Georgi Gerganov 2023-04-30 22:32:46 +0300
  • feac80dd3f
    ggml : fix UB (int << 31) Georgi Gerganov 2023-04-30 22:27:30 +0300
  • fa8dbdc888
    release : v1.4.0 v1.4.0 Georgi Gerganov 2023-04-30 19:23:37 +0300
  • 4a7d49af95
    examples : fix + refactor Levenshtein distance Georgi Gerganov 2023-04-30 19:12:49 +0300
  • 794b162a46
    whisper : add integer quantization support (#540) Georgi Gerganov 2023-04-30 18:51:57 +0300
  • 7aeb79dd49
    wasm : add links to other examples Georgi Gerganov 2023-04-30 18:49:39 +0300
  • d378f5d9be
    readme : add information for model quantization Georgi Gerganov 2023-04-30 18:33:37 +0300
  • ced04d0f13
    node : increase test timeout to 10s Georgi Gerganov 2023-04-30 18:29:49 +0300
  • 05137ceb84
    talk-llama : update example to latest llama.cpp Georgi Gerganov 2023-04-30 18:22:20 +0300
  • 7c437117c6
    wasm : bump DB version number to 2 Georgi Gerganov 2023-04-30 18:04:42 +0300
  • a3e96e78f6
    wasm : add quantized models to all WASM examples Georgi Gerganov 2023-04-30 17:54:19 +0300
  • 524470dcd2
    ggml : add Q5_0 and Q5_1 WASM SIMD Georgi Gerganov 2023-04-30 17:37:12 +0300
  • 9364cfd4d2
    bench.wasm : fix custom model button Georgi Gerganov 2023-04-30 15:36:21 +0300
  • ec3e882c86
    bench.wasm : remove "medium" button Georgi Gerganov 2023-04-30 15:34:23 +0300
  • f641252984
    wasm : update quantized models to Q5 Georgi Gerganov 2023-04-30 15:33:22 +0300
  • 23f63e0b63
    whisper : try to fix shared lib linkage Georgi Gerganov 2023-04-30 14:43:54 +0300
  • 8b91958822
    whisper : fix F32 support Georgi Gerganov 2023-04-30 14:33:49 +0300
  • 9d6bcad498
    whisper : quantization tool ready Georgi Gerganov 2023-04-30 14:28:12 +0300
  • d814e95e6d
    examples : add common-ggml + prepare to add "quantize" tool Georgi Gerganov 2023-04-30 13:39:12 +0300
  • e9d78e65d9
    whisper : add integer quantization support Georgi Gerganov 2023-04-30 12:57:16 +0300
  • 5fd1bdd7fc
    whisper : add GPU support via cuBLAS (#834) Georgi Gerganov 2023-04-30 12:14:33 +0300
  • ceb6570e1c
    gitignore : ignore build-cublas Georgi Gerganov 2023-04-30 12:13:44 +0300
  • 3c5dedb6e6
    cmake : add WHISPER_CUBLAS option Georgi Gerganov 2023-04-30 12:12:27 +0300
  • 4660596c40
    readme : add cuBLAS instructions Georgi Gerganov 2023-04-30 12:05:22 +0300
  • e29c57d372
    whisper : remove old commented code Georgi Gerganov 2023-04-30 11:58:46 +0300
  • 13431f3c19 whisper : disable Flash Attention + adjust memory buffers Georgi Gerganov 2023-04-30 11:49:00 +0300
  • c729b38064
    make : fix CUBLAS build Georgi Gerganov 2023-04-30 10:16:20 +0300
  • ae29ac0ad9 Update elevenlabs example to use ufficial python API DGdev91 2023-04-30 01:00:27 +0200
  • 582f559dbf
    make : add WHISPER_CUBLAS Georgi Gerganov 2023-04-29 19:35:21 +0300
  • 0ccd6746c9
    ggml : fix WASM build Georgi Gerganov 2023-04-29 21:37:23 +0300
  • d9b550c0a1
    ggml : fix 32-bit ARM NEON (#836) Georgi Gerganov 2023-04-29 21:33:33 +0300
  • ee60d7df0c
    ggml : fix Georgi Gerganov 2023-04-29 21:24:09 +0300
  • 6383346efa
    ggml : fix Georgi Gerganov 2023-04-29 21:20:24 +0300
  • 65c27873d6
    ggml : add support for 32-bit ARM Georgi Gerganov 2023-04-29 21:16:14 +0300
  • e9b091c92a
    ggml : use vzip instead of vuzp for consistency Georgi Gerganov 2023-04-29 21:14:09 +0300
  • cf33c1d0ef command: fix out of bounds access in levenshtein distance function Helmut 2023-04-29 20:09:27 +0200
  • 1f30b99208
    ggml : fix WASM build Georgi Gerganov 2023-04-29 20:21:25 +0300
  • 05c3ea3bc8
    ggml : sync with ggml repo (warning fixes + asserts) Georgi Gerganov 2023-04-29 19:30:22 +0300
  • 6108d3cc58
    whisper : use correct seek_end when offset is used (#833) Thijs Raymakers 2023-04-29 17:55:37 +0200
  • 20456470a0
    whisper: Use correct seek_end when offset is used Thijs Raymakers 2023-04-29 14:32:07 +0200
  • bab97c83d0
    tests : add "threads" to run-tests.sh Georgi Gerganov 2023-04-29 12:32:18 +0300
  • 3eaeb030ff
    extra : add sync-ggml.sh script Georgi Gerganov 2023-04-29 12:32:05 +0300
  • acec73ab6e
    ggml : sync latest ggml + llama.cpp updates (quantization) Georgi Gerganov 2023-04-29 12:31:52 +0300
  • 176c0dddee whisper.wasm : fix typo in readme Baffin Lee 2023-04-29 16:09:13 +0800
  • 5cc17418c7
    whisper.android : add some tips (#816) Zollner 2023-04-29 16:00:20 +0800
  • 3efb81dec6
    build : add WHISPER_COREML_ALLOW_FALLBACK to make / CMake (#812) Georgi Gerganov 2023-04-29 10:55:24 +0300
  • 94a7cd2a07
    whisper : allow non-CoreML fallback when Core ML cannot be loaded (#812) Canis Lupus 2023-04-29 08:49:02 +0100
  • 3e82ff4747
    whisper : fix bug from previous commit Georgi Gerganov 2023-04-29 10:42:14 +0300
  • b5bd2f43c5
    whisper : avoid designated initializers Georgi Gerganov 2023-04-29 10:36:50 +0300
  • 94aa56f19e
    minor : improve C++ and Python style (#768) AsukaMinato 2023-04-29 16:06:25 +0900
  • 0f73d655a6
    Merge 7a5a5fe86d into 4d89ee2e59 Jan Bielak 2023-04-28 15:41:34 -0400
  • 4d89ee2e59
    readme : add logo Georgi Gerganov 2023-04-28 22:41:29 +0300
  • ed64164bfe 2 space asukaminato 2023-04-24 03:41:30 +0900
  • 08f1c070cf Keep the original asukaminato 2023-04-24 03:40:07 +0900
  • 70567eff23
    main : escape quotes in csv output (#815) Laytan Laats 2023-04-23 18:01:59 +0200
  • 92644f7287 escape qoutes in csv output Laytan Laats 2023-04-23 03:14:09 +0200
  • 02ec83c5d5
    stream : flush upon finishing inference (#811) Taras Glek 2023-04-23 17:00:30 +0300
  • 2bd4b8d577
    examples : add missing #include <cstdint> (#798) Philipp Zabel 2023-04-23 15:52:52 +0200
  • eecf2c3d41
    main : update escape_double_quotes() function (#776) Tauseef Mohiuddin 2023-04-23 08:47:30 -0500
  • 1de07f7d5e
    add on some tips about in the readme of the android project folder Zollner 2023-04-23 15:10:47 +0800
  • 318f9ed2e4 When building with WHISPER_USE_COREML and WHISPER_COREML_ALLOW_FALLBACK, if the Core ML model cannot be loaded, continue without Core ML instead of returning. This allows a single build to transcribe using Core ML models where available, and regular models when not. Canis Lupus 2023-04-21 17:58:12 +0100
  • 57051d2f1f Flush upon finishing inference Taras Glek 2023-04-21 04:29:21 +0300
  • 0c797ffafc examples : add missing #include <cstdint> Philipp Zabel 2023-04-19 07:48:22 +0200
  • 7a5a5fe86d
    Merge branch 'ggerganov:master' into improve-performance Jan Bielak 2023-04-17 10:31:29 +0200
  • 3ad7a52607 Updated escape_double_quotes() Function tauseefmohammed2 2023-04-16 21:48:55 -0500
  • e9c3ca7e7a const some iter asukaminato 2023-04-16 06:48:36 +0900
  • b0eec2f96d recover some format asukaminato 2023-04-16 06:35:51 +0900
  • 827a5a7491 use self.field than setattr, use pathlib.Path asukaminato 2023-04-16 06:21:36 +0900
  • 55cd1f8f6b use some STL functions asukaminato 2023-04-16 05:15:03 +0900
  • c23588cc4b
    release : v1.3.0 v1.3.0 Georgi Gerganov 2023-04-15 17:30:44 +0300
  • 5108b30e6d
    whisper : pad audio instead of spectrogram (#579) Georgi Gerganov 2023-04-15 17:18:43 +0300
  • f19e23fbd1
    whisper : restore decoder temperature fallbacks Georgi Gerganov 2023-04-15 16:04:07 +0300