Commit Graph

  • 935d401ac7
    Merge dac938391d into f92bd59951 Greener-Dalii 2025-03-30 09:14:13 +0200
  • aaab74ea08
    Merge 5c76377b09 into f92bd59951 Ranjit 2025-03-30 06:08:41 +0200
  • ada1d14a24
    Merge 960d7e6d71 into f92bd59951 Jorropo 2025-03-30 06:06:31 +0200
  • 2d1e50f0bc
    Merge d29493d14e into f92bd59951 philipag 2025-03-30 14:00:18 +1000
  • c0588a226d
    Merge 524140c8fb into f92bd59951 Amanda Der Bedrosian 2025-03-30 05:56:21 +0200
  • f92bd59951
    whisper : remove unnecessary GGML_UNUSED macro (#2960) master b2349 Daniel Bevenius 2025-03-30 05:56:10 +0200
  • 2680d8ce70
    Merge 613f938ecf into 6e7629b146 thewh1teagle 2025-03-29 11:49:29 +0100
  • 058a9264d4
    Merge d0f38def08 into 6e7629b146 Tamotsu Takahashi 2025-03-28 17:29:45 -0400
  • 125711256c
    Merge 82f84e5b2e into 6e7629b146 Lars Fernhomberg 2025-03-28 17:29:44 -0400
  • f4ad18dd0b
    Merge 6c05cf7b0d into 6e7629b146 Bjarke Viksøe 2025-03-28 17:29:44 -0400
  • 61c82f35b0
    Merge ec05d7705a into 6e7629b146 wznmickey 2025-03-28 17:29:44 -0400
  • 4b0caa59de
    Merge 3972a8c6a5 into 6e7629b146 kristianmk 2025-03-28 17:29:44 -0400
  • fee52f62e5
    Merge 339704be07 into 6e7629b146 Thomas Fitzsimmons 2025-03-28 17:29:44 -0400
  • d3f2c68a73
    Merge 70f0b4a587 into 6e7629b146 Daniel Bevenius 2025-03-28 17:29:44 -0400
  • 61d9ea0f11
    Merge 2e1fb518d1 into 6e7629b146 DrEmixam 2025-03-28 17:29:43 -0400
  • e9b76cb34a
    Merge 4c60e6b0c1 into 6e7629b146 Andreas Lubbe 2025-03-28 17:29:43 -0400
  • 6e7629b146 sync : ggml b2348 Georgi Gerganov 2025-03-28 20:58:21 +0200
  • 27533e7f63 metal : improve FA + improve MoE (llama/12612) Georgi Gerganov 2025-03-28 20:21:59 +0200
  • 1b81415963 vulkan: fix coopmat shader generation when cross-compiling (llama/12272) Icenowy Zheng 2025-03-29 01:51:06 +0800
  • 0001ec075f llamafile : ppc64le GEMV forwarding for FP32. (llama/12594) amritahs-ibm 2025-03-28 13:13:22 +0530
  • 5bad2e5099 rpc : send hash when tensor data is above some fixed threshold (llama/12496) Radoslav Gerganov 2025-03-28 08:18:04 +0200
  • 6fc0ae2f5a opencl: add multi and vision rope, gelu_quick and im2col (llama/12600) lhez 2025-03-27 08:08:08 -0700
  • 8ba897b7f5
    sync : ggml Georgi Gerganov 2025-03-28 20:58:21 +0200
  • 0e03fc9c23
    metal : improve FA + improve MoE (llama/12612) Georgi Gerganov 2025-03-28 20:21:59 +0200
  • feea2011f2
    vulkan: fix coopmat shader generation when cross-compiling (llama/12272) Icenowy Zheng 2025-03-29 01:51:06 +0800
  • 3141956360
    llamafile : ppc64le GEMV forwarding for FP32. (llama/12594) amritahs-ibm 2025-03-28 13:13:22 +0530
  • 7b8090810e
    rpc : send hash when tensor data is above some fixed threshold (llama/12496) Radoslav Gerganov 2025-03-28 08:18:04 +0200
  • 263a5888d3
    opencl: add multi and vision rope, gelu_quick and im2col (llama/12600) lhez 2025-03-27 08:08:08 -0700
  • c254ec6f95 whisper : remove unnecessary GGML_UNUSED macro Daniel Bevenius 2025-03-28 12:37:49 +0100
  • de6b38c6d9
    bindings.go : add DetectedLanguage to go bindings (#2947) b2342 Amanda Der Bedrosian 2025-03-28 04:26:22 -0700
  • 7c5146ddb0 Revert "remove : dummy commit to trigger ci" [no ci] Daniel Bevenius 2025-03-28 12:09:48 +0100
  • f657f22106 remove : dummy commit to trigger ci Daniel Bevenius 2025-03-28 11:40:34 +0100
  • 54454480fe ci : add src and include to paths filter Daniel Bevenius 2025-03-28 11:38:44 +0100
  • c892cef63d ci : add build of wasm examples to CI Daniel Bevenius 2025-03-28 10:45:18 +0100
  • 46d6e0abc1
    ruby : fix test failures in test_whisper (#2955) b2341 Daniel Bevenius 2025-03-28 09:29:56 +0100
  • 70f0b4a587 ci : re-enable android_java job Daniel Bevenius 2025-03-28 09:08:37 +0100
  • 5f75cae0b5 ci : fix whisper.dll path in build.yml Daniel Bevenius 2025-03-28 08:48:16 +0100
  • 4c0c912176 ci : use arch for .dll names and enable jna debug Daniel Bevenius 2025-03-28 08:38:19 +0100
  • fa8c577b14 ci : fix List build release files step Daniel Bevenius 2025-03-28 08:08:01 +0100
  • 956ceefd58 ci : fix copy of whiper.ddl to build\Release dir Daniel Bevenius 2025-03-28 07:53:42 +0100
  • 1279f0d0bc
    examples : support progress_callback API for addon.node (#2941) b2340 Lin Xiaodong 2025-03-28 13:34:26 +0800
  • 36fa375b81 ci : add BUILD_SHARED_LIBS=ON windows build option Daniel Bevenius 2025-03-27 19:59:10 +0100
  • 14ffc5e282 ci : copy SDL2.dll to build\Release\SDL2.dll Daniel Bevenius 2025-03-27 19:27:53 +0100
  • fdeea64b86 ci : fix path to SDL2.dll Daniel Bevenius 2025-03-27 19:01:56 +0100
  • 95288a8f99 ci : fix sdl2.dll upload and download Daniel Bevenius 2025-03-27 18:50:20 +0100
  • 2982bf72bb ci : move SDL2.dll upload to correct job Daniel Bevenius 2025-03-27 18:09:58 +0100
  • 1b76698c9c ci : download SDL2.dll and copy it to the resources directory Daniel Bevenius 2025-03-27 17:20:34 +0100
  • f3c9030875 ci : add logging to debug JNA library loading Daniel Bevenius 2025-03-27 16:37:11 +0100
  • 70f35b186d bindings.java : update destination path for native libraries Daniel Bevenius 2025-03-27 15:57:53 +0100
  • 8b1661a667 ci : try copying the DLL to build/Release Daniel Bevenius 2025-03-27 15:39:08 +0100
  • 4f9a7dbb9b ci: move .dll to correct location bindings-java Daniel Bevenius 2025-03-27 15:00:34 +0100
  • 7129bbfed9 squash! ci : re-enable bindings-java (java) job Daniel Bevenius 2025-03-27 14:37:29 +0100
  • bfc213d2d0 squash! ci : re-enable bindings-java (java) job Daniel Bevenius 2025-03-27 14:02:03 +0100
  • 5b141a977e squash! ci : re-enable bindings-java (java) job Daniel Bevenius 2025-03-27 13:38:43 +0100
  • 0208803b66 ci : re-enable bindings-java (java) job Daniel Bevenius 2025-03-27 13:27:35 +0100
  • 3660e14588 bindings.ruby : enable Wisper.log_set in tests Daniel Bevenius 2025-03-27 12:55:43 +0100
  • fa28968800 bindings.ruby : fix warnings in tests Daniel Bevenius 2025-03-27 06:36:14 +0100
  • 41c6e25bca bindings.ruby : fix test failures in test_whisper Daniel Bevenius 2025-03-27 10:53:46 +0100
  • f28bf5d186 xcf : fix visionOS build b2339 Georgi Gerganov 2025-03-27 10:30:09 +0200
  • 1fbdfb1d36 files : remove old wkv6 (#0) Georgi Gerganov 2025-03-27 10:15:02 +0200
  • ee5581633b sync : ggml Georgi Gerganov 2025-03-27 10:13:47 +0200
  • 8ca67df291 ggml : sync/merge cmake,riscv,powerpc, add common.cmake (ggml/0) Georgi Gerganov 2025-03-27 09:12:54 +0200
  • fc6d343e76 llamafile : ppc64le MMA implementation for Q4_0. (llama/12489) amritahs-ibm 2025-03-27 12:21:47 +0530
  • 3199356d3a SYCL: implement memset ggml backend buffer interface (llama/12580) Akarshan Biswas 2025-03-27 07:16:00 +0530
  • e0c43b0bbf HIP: Add support for RDNA4 targets (llama/12372) Slobodan Josic 2025-03-26 23:46:30 +0100
  • f4f619ea8e metal : refactor mat-vec code (llama/12569) Georgi Gerganov 2025-03-26 21:38:38 +0200
  • 3c4d363872 ggml : fix MUL_MAT_ID repack with Q8_K (llama/12544) Georgi Gerganov 2025-03-26 13:02:00 +0200
  • 15aa189329 ggml-cpu : update KleidiAI to v1.5.0 (llama/12568) Dan Johansson 2025-03-25 12:10:18 +0100
  • c53d5c9e85 SYCL: disable Q4_0 reorder optimization (llama/12560) Akarshan Biswas 2025-03-25 16:10:18 +0530
  • ba6f584f30 opencl: simplify kernel embedding logic in cmakefile (llama/12503) lhez 2025-03-24 09:20:47 -0700
  • a219941812 CUDA: Fix clang warnings (llama/12540) R0CKSTAR 2025-03-24 18:28:34 +0800
  • a2cc8c2666 vulkan: fix mul_mat_vec failure in backend tests (llama/12529) Jeff Bolz 2025-03-24 01:56:17 -0500
  • 388ed98220 ggml : fix quantized cpy op (llama/12310) Georgi Gerganov 2025-03-22 16:23:26 +0200
  • d487a28ae1 musa: refine compute capability (llama/12493) R0CKSTAR 2025-03-22 17:11:37 +0800
  • cbb88c4050 vulkan: Optimize mul_mat_vec p021 and nc shaders (llama/12505) Jeff Bolz 2025-03-22 03:40:11 -0500
  • 13455c0b5f Vulkan: RTE rounding for cpy to quant (llama/12480) stduhpf 2025-03-21 20:34:50 +0100
  • 2f77a9e9bd vulkan: workaround for AMD Windows driver 16 bit unpack8 bug (llama/12472) Eve 2025-03-21 19:27:47 +0000
  • fa2b5249ff Fix build on Windows when ccache enabled (ggml/9954) (llama/9976) 蕭澧邦 2025-03-21 14:58:47 +0800
  • 5b854ebba5 sycl: cleanup oneDNN related code (llama/12097) Svetlozar Georgiev 2025-03-21 02:15:56 +0000
  • 8058f19d0b ggml : block interleaving support for Q4_K quantization for x86 AVX2 architecture (llama/12332) Srihari-mcw 2025-03-20 17:05:34 +0530
  • ae6a9bb9a5 CUDA: Improve flash decoding kernel GPU occupancy for BS=1 case (llama/12183) Gaurav Garg 2025-03-20 01:22:06 +0530
  • 24faba9e9b vulkan: optimize iq1 coopmat2 dequant functions (llama/12427) Jeff Bolz 2025-03-19 13:56:23 -0500
  • c722ff84d3 Fix visionOS build and add CI (llama/12415) Guus Waals 2025-03-19 10:15:23 +0000
  • 102af79f63 vulkan: Submit once enough matmul work has been recorded (llama/12406) Jeff Bolz 2025-03-19 02:26:26 -0500
  • 03c364557d opencl: improve profiling (llama/12442) lhez 2025-03-18 12:54:55 -0700
  • 31b62276cf musa: override warp_size of musa device to 32 (llama/12445) R0CKSTAR 2025-03-19 02:28:26 +0800
  • 97b5a3055d SYCL: using graphs is configurable by environment variable and compile option (llama/12371) Łukasz Ślusarczyk 2025-03-18 11:16:31 +0100
  • 9993c3f703 ggml : add SVE support for q6_K_q8_K (llama/12361) fj-y-saito 2025-03-18 17:14:39 +0900
  • fa72479cfb Vulkan: Default to 1GB allocations instead of 4GB to avoid fragmentation and driver issues (llama/12434) 0cc4m 2025-03-18 07:21:40 +0100
  • 6c15539c54 fixed compilation warnings in ggml-sycl (llama/12424) Łukasz Ślusarczyk 2025-03-18 01:51:25 +0100
  • 52c4c03b0a llama: Add support for RWKV v7 architecture (llama/12412) Molly Sophia 2025-03-18 07:27:50 +0800
  • cfc2560e41 cuda : enable CUDA Graph on CUDA Toolkit < 12.x (llama/12394) Gaurav Garg 2025-03-17 23:55:13 +0530
  • db6e8056b5 ggml-vulkan: remove unused find_program(glslc) (llama/12416) Guus Waals 2025-03-18 00:35:43 +0800
  • b3f3779c1b vulkan: Add N/2 and N/4 optimized paths in coopmat2 shader (llama/12312) Jeff Bolz 2025-03-17 09:26:18 -0500
  • 13eeebb1b2 vulkan: subgroup size tuning (llama/12087) Daniele 2025-03-17 12:42:33 +0100
  • 905b834af1 vulkan: use fp32 in coopmat2 q4_k dequant function (llama/12309) Jeff Bolz 2025-03-17 04:43:35 -0500
  • 2cd3061a23 vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bounds checking (llama/12273) Jeff Bolz 2025-03-17 04:41:59 -0500
  • 88d59e21b2 vulkan: Adjust coopmat2 tile sizes and selection heuristic (llama/12258) Jeff Bolz 2025-03-17 04:35:00 -0500
  • 4917f122d4 cmake : enable building llama.cpp using system libggml (llama/12321) Christian Kastner 2025-03-17 10:05:23 +0100
  • 16a1b77249 SYCL: set extras only on GGML_TYPE_Q4_0 (llama/12366) Akarshan Biswas 2025-03-17 07:15:12 +0530