Commit Graph

  • 4e89677370
    Merge dac938391d into 3de9deead5 #2272 Greener-Dalii 2024-12-18 18:39:20 +0100
  • 422647fcbe
    Merge 8d60666cbc into 3de9deead5 #2291 Amanda Der Bedrosian 2024-12-18 18:39:20 +0100
  • b936e0152e
    Merge 613f938ecf into 3de9deead5 #2369 thewh1teagle 2024-12-18 18:39:20 +0100
  • 4b67c04a70
    Merge 2651c3e2d4 into 3de9deead5 #2384 shivghai 2024-12-18 18:39:20 +0100
  • 353755c698
    Merge 69b3136171 into 3de9deead5 #2406 Dave Lewis 2024-12-18 18:39:20 +0100
  • 159f2d604f
    Merge b87d6691f5 into 3de9deead5 #2560 Pranav Seelam 2024-12-18 18:39:20 +0100
  • 70ae32b55d
    Merge c5b9b546b8 into 3de9deead5 #2569 Tomer Schlesinger 2024-12-18 18:39:20 +0100
  • 323e5dc604
    Merge 60c293e943 into 3de9deead5 #2577 Karthick 2024-12-18 18:39:20 +0100
  • 4bcfecaa57
    Merge b0aeef2d52 into 3de9deead5 #2579 Georgi Gerganov 2024-12-18 18:39:20 +0100
  • 4a0db5551c
    Merge 6c05cf7b0d into 3de9deead5 #2279 Bjarke Viksøe 2024-12-18 19:15:18 +0300
  • 2869b58f8b
    Merge 2e1fb518d1 into 3de9deead5 #2593 DrEmixam 2024-12-18 10:13:40 -0600
  • 724cd1ff17
    Merge ec05d7705a into 3de9deead5 #2634 wznmickey 2024-12-18 16:13:22 +0000
  • 3de9deead5
    release : v1.7.3 master v1.7.3 Georgi Gerganov 2024-12-18 18:12:40 +0200
  • 47f989f9b3
    ci : msys enable SDL2 build (#2635) Georgi Gerganov 2024-12-18 12:52:41 +0200
  • acc4e13dee ruby : sync ggml (#2643) KITAITI Makoto 2024-12-18 19:51:48 +0900
  • ba6c2a8fd9 android : try to fix build Georgi Gerganov 2024-12-18 09:54:23 +0200
  • 6576af00d7 files : remove old sources Georgi Gerganov 2024-12-18 08:39:39 +0200
  • 8ac5db0169 sync : ggml Georgi Gerganov 2024-12-18 08:38:56 +0200
  • 61edb117a0 talk-llama : sync llama.cpp Georgi Gerganov 2024-12-17 21:19:28 +0200
  • eb97b257eb sync : ggml Georgi Gerganov 2024-12-17 19:25:44 +0200
  • 479499dc0e ggml : update ggml_backend_cpu_device_supports_op (llama/10867) Georgi Gerganov 2024-12-17 18:35:42 +0200
  • d420a759c5 vulkan: bugfixes for small subgroup size systems + llvmpipe test (llama/10809) Eve 2024-12-17 05:52:55 +0000
  • a1ab9b5e91 rwkv6: add wkv6 support for Vulkan backend (llama/10829) Zhiyuan Li 2024-12-17 05:00:46 +0800
  • e22d38e4f2 llama : add Qwen2VL support + multimodal RoPE (llama/10361) HimariO 2024-12-14 20:43:46 +0800
  • 856fbaa92f Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (llama/10693) lhez 2024-12-13 12:23:52 -0800
  • 2c05efa4b1 Fix crash caused by ggml_backend_load_all when launching on Android Activity (llama/10812) 谢乃闻 2024-12-13 12:56:07 +0000
  • c21fb10b28 vulkan: small mul_mat_vec optimizations (llama/10665) Eve 2024-12-13 08:42:04 +0000
  • 26c9fd0cdc SYCL: Reduce most of the compiler warnings (llama/10748) Akarshan Biswas 2024-12-13 12:12:15 +0530
  • e6eed605cf ggml : Fix compilation issues on ARM platform when building without fp16 (llama/10811) Karol Kontny 2024-12-13 01:04:19 +0100
  • abe3102cb7 CUDA: faster non-contiguous concat (llama/10760) a3sh 2024-12-13 02:09:50 +0800
  • 1193e494a9 remove CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS (llama/10797) Diego Devesa 2024-12-12 19:02:49 +0100
  • e5e951672e Vulkan: Use improved q4_k and q5_k dequant code in dequant shaders (llama/10798) 0cc4m 2024-12-12 18:36:00 +0100
  • 0e24559ad9 Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats (llama/10721) 0cc4m 2024-12-12 18:35:37 +0100
  • 527ac800cf ggml: load all backends from a user-provided search path (llama/10699) Gilad S 2024-12-11 02:47:21 +0200
  • 479bd77169 vulkan: request round-to-even for fp16 in im2col/rope_head (llama/10767) Jeff Bolz 2024-12-10 14:23:17 -0600
  • d8bf63a41b vulkan: dynamic subgroup size for the remaining k quants (llama/10745) Eve 2024-12-10 19:33:23 +0000
  • b82c8d76dc CUDA: rename macros to avoid conflicts with WinAPI (llama/10736) Andreas Kieslinger 2024-12-10 18:23:24 +0100
  • 86346f811e vulkan: disable spirv-opt for coopmat shaders (llama/10763) Jeff Bolz 2024-12-10 11:22:20 -0600
  • c635f40a34 ggml : remove return from ggml_gallocr_allocate_node (ggml/1048) Daniel Bevenius 2024-12-14 03:23:08 +0100
  • e0be0de1ee ggml : add check for grad_accs (ggml/1046) Daniel Bevenius 2024-12-13 08:19:38 +0100
  • 60dc6d003f common : remove old types Georgi Gerganov 2024-12-10 17:19:09 +0200
  • eb27e0d834 CUDA: fix shared memory access condition for mmv (llama/10740) Johannes Gäßler 2024-12-09 20:07:12 +0100
  • a682fdce0c vulkan: fix compile warnings (llama/10731) Jeff Bolz 2024-12-09 01:24:01 -0600
  • 9ffbd3d969 Vulkan: fix NaN in tanh.comp with AMD proprietary driver on Windows (llama/10723) stduhpf 2024-12-08 19:19:19 +0100
  • 6585a890b4 vulkan: compile a test shader in cmake to check for coopmat2 support (llama/10713) Jeff Bolz 2024-12-08 02:05:55 -0600
  • d0a050b51f ggml : disable iq4_nl interleave size 8 (llama/10709) Georgi Gerganov 2024-12-07 18:38:15 +0200
  • e990d1b791 ggml : refactor online repacking (llama/10446) Djip007 2024-12-07 13:37:50 +0100
  • 4a6d52efe6 Vulkan: VK_KHR_cooperative_matrix support to speed up prompt processing (llama/10597) 0cc4m 2024-12-07 10:24:15 +0100
  • 8b841d430a metal : Extend how Llama.cpp locates metal resources (llama/10676) Robert Ormandi 2024-12-07 01:55:01 -0600
  • b74b68212a vulkan: Add VK_NV_cooperative_matrix2 support for mul_mat and flash attention (llama/10206) Jeff Bolz 2024-12-05 13:15:05 -0600
  • b145e9bf9c
    ruby : sync ggml (#2643) #2639 KITAITI Makoto 2024-12-18 19:51:48 +0900
  • bb62f4241f ruby : sync ggml #2643 Kitaiti Makoto 2024-12-18 19:28:50 +0900
  • 3a27b2b91b
    ruby : Add no_speech_thold (#2641) KITAITI Makoto 2024-12-18 18:00:50 +0900
  • f311d82e91
    android : try to fix build Georgi Gerganov 2024-12-18 09:54:23 +0200
  • d34445e960
    stream : improve consistency in README (#2642) crummyh 2024-12-18 00:43:48 -0600
  • 69fe4b7c7d
    files : remove old sources Georgi Gerganov 2024-12-18 08:39:39 +0200
  • f18f7c64c4
    sync : ggml Georgi Gerganov 2024-12-18 08:38:56 +0200
  • 161f4f6a9f
    Improve consistency in stream README #2642 crummyh 2024-12-17 16:41:24 -0600
  • 4d9b7365c3 Add test for Whisper::Params#no_speech_thold #2641 Kitaiti Makoto 2024-12-18 05:53:22 +0900
  • ca1127410c Add Whisper::Params#no_speech_thold attribute Kitaiti Makoto 2024-12-18 05:53:04 +0900
  • 218f22ea5f Remove redundant code from README for simplicity Kitaiti Makoto 2024-12-12 23:57:35 +0900
  • a71b8ff11f Initialize Whisper::Params#diarize explicitely Kitaiti Makoto 2024-12-12 23:47:17 +0900
  • b89d3972b0 Remove unnecessary flags Kitaiti Makoto 2024-12-12 08:05:27 +0900
  • fb4ec92d56 Fix a typo Kitaiti Makoto 2024-12-12 07:43:16 +0900
  • ede026d0b4 Specify required Ruby version Kitaiti Makoto 2024-12-10 07:22:25 +0900
  • 4f632dd6d6 Add test for auto download Kitaiti Makoto 2024-12-10 06:33:02 +0900
  • 3f628b149b Use cache model file if download fails Kitaiti Makoto 2024-12-10 06:32:54 +0900
  • dec133e445 Pass String to raise Kitaiti Makoto 2024-12-10 05:59:35 +0900
  • 36e84da86a Don't show download progress when not tty Kitaiti Makoto 2024-12-10 05:53:31 +0900
  • bffe4a1180 Update document comment of Whisper::Context#initialize Kitaiti Makoto 2024-12-10 05:47:40 +0900
  • 71b700b1a6 Move whisper/model.rb -> whisper/model/uri.rb Kitaiti Makoto 2024-12-10 05:46:02 +0900
  • 6960117938 Remove unnecessary task Kitaiti Makoto 2024-12-10 05:43:29 +0900
  • b8ff45059a Update README Kitaiti Makoto 2024-12-10 05:39:30 +0900
  • cde4b581a0 Use downloading pre-converted model feature for testing Kitaiti Makoto 2024-12-10 05:35:27 +0900
  • 241db0c987 Make Whisper::Context#initialize accept pre-converted model name Kitaiti Makoto 2024-12-10 05:34:39 +0900
  • f99bc8aca7 Fix Whisper::Model::URI#request Kitaiti Makoto 2024-12-10 05:26:54 +0900
  • 6c780b5f71 Remove Whisper::Model.[] Kitaiti Makoto 2024-12-10 04:57:25 +0900
  • 2add8c0ad8
    talk-llama : sync llama.cpp Georgi Gerganov 2024-12-17 21:19:28 +0200
  • 59494c03f4
    sync : ggml Georgi Gerganov 2024-12-17 19:25:44 +0200
  • 6b07615e35
    ggml : update ggml_backend_cpu_device_supports_op (llama/10867) Georgi Gerganov 2024-12-17 18:35:42 +0200
  • e7f9dbbbde
    vulkan: bugfixes for small subgroup size systems + llvmpipe test (llama/10809) Eve 2024-12-17 05:52:55 +0000
  • a26ce8b03e
    rwkv6: add wkv6 support for Vulkan backend (llama/10829) Zhiyuan Li 2024-12-17 05:00:46 +0800
  • 6a3b05a23d
    llama : add Qwen2VL support + multimodal RoPE (llama/10361) HimariO 2024-12-14 20:43:46 +0800
  • 5538f04772
    Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (llama/10693) lhez 2024-12-13 12:23:52 -0800
  • 874fcb62cc
    Fix crash caused by ggml_backend_load_all when launching on Android Activity (llama/10812) 谢乃闻 2024-12-13 12:56:07 +0000
  • dcef9a4fdb
    vulkan: small mul_mat_vec optimizations (llama/10665) Eve 2024-12-13 08:42:04 +0000
  • d1abb1e824
    SYCL: Reduce most of the compiler warnings (llama/10748) Akarshan Biswas 2024-12-13 12:12:15 +0530
  • 3ead635fc4
    ggml : Fix compilation issues on ARM platform when building without fp16 (llama/10811) Karol Kontny 2024-12-13 01:04:19 +0100
  • 88b83a3688
    CUDA: faster non-contiguous concat (llama/10760) a3sh 2024-12-13 02:09:50 +0800
  • b79bf1c757
    remove CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS (llama/10797) Diego Devesa 2024-12-12 19:02:49 +0100
  • 949dd3d236
    Vulkan: Use improved q4_k and q5_k dequant code in dequant shaders (llama/10798) 0cc4m 2024-12-12 18:36:00 +0100
  • 049e991f52
    Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats (llama/10721) 0cc4m 2024-12-12 18:35:37 +0100
  • cfeb7f707e
    ggml: load all backends from a user-provided search path (llama/10699) Gilad S 2024-12-11 02:47:21 +0200
  • 8368c1dcfc
    vulkan: request round-to-even for fp16 in im2col/rope_head (llama/10767) Jeff Bolz 2024-12-10 14:23:17 -0600
  • e67025526d
    vulkan: dynamic subgroup size for the remaining k quants (llama/10745) Eve 2024-12-10 19:33:23 +0000
  • 5fd6abb861
    CUDA: rename macros to avoid conflicts with WinAPI (llama/10736) Andreas Kieslinger 2024-12-10 18:23:24 +0100
  • d241e1b254
    vulkan: disable spirv-opt for coopmat shaders (llama/10763) Jeff Bolz 2024-12-10 11:22:20 -0600
  • a619ddfa63
    ggml : remove return from ggml_gallocr_allocate_node (ggml/1048) Daniel Bevenius 2024-12-14 03:23:08 +0100
  • ffda777993
    ggml : add check for grad_accs (ggml/1046) Daniel Bevenius 2024-12-13 08:19:38 +0100
  • 8541e2cf9e
    common : remove old types Georgi Gerganov 2024-12-10 17:19:09 +0200