Commit Graph

  • 77d708fabb rpc : set SO_REUSEADDR for the server socket (llama/7320) Radoslav Gerganov 2024-05-17 17:25:44 +03:00
  • c00149c861 ggml-quants, llama : removed excess checks (llama/7274) Herman Semenov 2024-05-17 07:08:49 +00:00
  • 574661f2e6 ggml : rewrite silu and softmax for cpu (llama/7154) Justine Tunney 2024-05-17 02:58:52 -04:00
  • 7bd69349bf rpc : add command line arg for specifying backend memory Radoslav Gerganov 2024-05-15 15:29:07 +03:00
  • 488ad99c13 Add support for properly optimized Windows ARM64 builds with LLVM and MSVC (llama/7191) Max Krasnyansky 2024-05-15 19:47:36 -07:00
  • 7178cceeaa ggml : use dynamic thread scheduling for matrix multiplication (llama/6915) kunnis 2024-05-15 12:59:12 -05:00
  • 8d55ccdb8c Avoid unnecessarily disabling CUDA graphs (llama/7302) agray3 2024-05-15 14:44:49 +01:00
  • 37a72cb170 ggml : tag ggml_tensor::backend as deprecated (llama/7290) slaren 2024-05-15 15:08:48 +02:00
  • bf9b69284f Add missing " (llama/7303) AidanBeltonS 2024-05-15 13:26:30 +01:00
  • c4de1e19df ggml : add ggml_upscale_ext (ggml/814) John Balis 2024-05-15 03:52:33 -05:00
  • 5b7073cae1 scripts : update sync Georgi Gerganov 2024-06-16 12:41:42 +03:00
  • b29b3b2924 whisper : use ggml-cuda in mel calc, set appropriate device (#2236) Borislav Stanimirov 2024-06-13 13:16:07 +03:00
  • 267e15a46d cuda : avoid async allocs in CUDA mel code gg/cuda-no-async Georgi Gerganov 2024-06-12 09:52:15 +03:00
  • 420b6abc54 cuda : fix HIPBLAS build (#2234) Georgi Gerganov 2024-06-11 19:14:38 +03:00
  • 5801b8ac64 cuda : fix HIPBLAS build gg/hipblas-fix Georgi Gerganov 2024-06-11 19:13:43 +03:00
  • 99804b0f3e cuda : fix bounds check for src0 rows in MMVQ kernel (#2231) Georgi Gerganov 2024-06-11 17:39:01 +03:00
  • 13c5446759 Update ggml-cuda/mmvq.cu gg/cuda-fix-mmvq Georgi Gerganov 2024-06-11 17:37:32 +03:00
  • c55964c956 ci : fix CUDA builds (#2232) Georgi Gerganov 2024-06-11 17:21:30 +03:00
  • 059bcd3009 ci : fix CUDA builds gg/ci-cuda-fix Georgi Gerganov 2024-06-11 11:40:19 +03:00
  • 9df6298a91 cuda : fix bounds check for src0 rows in MMVQ kernel Georgi Gerganov 2024-06-11 11:30:12 +03:00
  • 20c542c713 whisper : auto-grow working areas for mel_calc_cuda (#2227) Borislav Stanimirov 2024-06-10 21:51:32 +03:00
  • c2bdb960cd whisper : free whisper_mel instances (#2220) Georgi Gerganov 2024-06-10 10:59:36 +03:00
  • 87acd6d629 whisper : whisper_state/backend fixes (#2217) Georgi Gerganov 2024-06-06 18:51:36 +03:00
  • f842d31171 whisper : calculate mel spectrogram directly into a ggml_tensor (#2208) Borislav Stanimirov 2024-06-06 16:20:46 +03:00
  • ffef323c4c whisper : add CUDA-specific computation mel spectrograms (#2206) Borislav Stanimirov 2024-06-04 09:32:23 +03:00
  • af5833e298 whisper : remove speed_up and phase_vocoder* functions (#2198) Borislav Stanimirov 2024-05-31 11:37:29 +03:00
  • b87494bb8f readme : add conan badge (#2196) Martin Delille 2024-05-30 14:43:28 +02:00
  • ad130431aa readme : add install instructions for Conan (#2189) Carlos Zoido 2024-05-30 14:06:15 +02:00
  • e130b66642 whisper: use global cache for sin/cos vals and Hann window (#2194) Borislav Stanimirov 2024-05-29 19:09:21 +03:00
  • c7b6988678 release : v1.6.2 v1.6.2 Georgi Gerganov 2024-05-27 10:35:09 +03:00
  • 05042a782d Revert "whisper : remove extra backend instance (huh?)" (#2182) Georgi Gerganov 2024-05-27 10:20:25 +03:00
  • a7dc2aab16 server : fix typo (#2181) Daniel Valdivia 2024-05-25 00:46:22 -07:00
  • 22d46b7ba4 ruby : update bindings (#2154) Todd 2024-05-22 16:02:52 -04:00
  • c10db6ea28 release : v1.6.1 v1.6.1 Georgi Gerganov 2024-05-21 18:44:37 +03:00
  • 1b51fdf170 examples : add support for decoding input with ffmpeg (Linux) (#2133) William Tambellini 2024-05-21 08:31:41 -07:00
  • adee3f9c1f node : add flash_attn param (#2170) Pedro Probst 2024-05-20 03:08:48 -03:00
  • 4798be1f9a ci: Update build.yml to suppress warnings about node.js versions (#2166) Tamotsu Takahashi 2024-05-19 17:49:26 +09:00
  • 08981d1bac release : v1.6.0 v1.6.0 Georgi Gerganov 2024-05-15 09:59:48 +03:00
  • 7094ea5e75 whisper : use flash attention (#2152) Georgi Gerganov 2024-05-15 09:38:19 +03:00
  • 9d5771ae43 talk-llama : reject runs without required arguments (#2153) petterreinholdtsen 2024-05-14 20:32:41 +02:00
  • f56b8305c4 sync : ggml Georgi Gerganov 2024-05-14 19:16:32 +03:00
  • 1056ad762c metal : support FA without mask + add asserts (llama/7278) Georgi Gerganov 2024-05-14 19:09:30 +03:00
  • c451080c8b ggml : add RPC backend (llama/6829) Radoslav Gerganov 2024-05-14 14:27:19 +03:00
  • 8e7c22fbdb rm wait() (llama/7233) Neo Zhang 2024-05-13 18:11:26 +08:00
  • e57e95eb0d CUDA: add FP32 FlashAttention vector kernel (llama/7188) Johannes Gäßler 2024-05-12 19:40:45 +02:00
  • 130f43e4b8 scripts : sync ggml-rpc Georgi Gerganov 2024-05-14 19:15:35 +03:00
  • d8356a1cc2 whisper : fix model path encoding in windows (#2086) thewh1teagle 2024-05-14 09:43:41 +03:00
  • 4ef8d9f44e server : return utf-8 (#2138) Georgi Gerganov 2024-05-13 15:33:46 +03:00
  • 3928dbd206 node : add audio_ctx and audio buffer params (#2123) Pedro Probst 2024-05-13 09:22:23 -03:00
  • 2ced6f0742 cmake : fix HIP/ROCm build (#2102) aldorof 2024-05-13 08:18:43 -04:00
  • 30f73109b8 node : add additional params (#2000) valVk 2024-05-13 15:15:43 +03:00
  • 17fa62d3d3 js : remove un-needed request header from fetchRemote (#2119) Mark Karpelès 2024-05-13 21:13:19 +09:00
  • 1da5edcde0 cmake : fix metal embed sources path (#2110) Georgi Gerganov 2024-05-13 15:09:35 +03:00
  • 0bb05b113d main : dont print timings with --no-prints (#2108) Daniel Ziegenberg 2024-05-13 14:00:19 +02:00
  • f141b2b938 main : add options for temperature control (#2088) Daniel Ziegenberg 2024-05-13 13:59:44 +02:00
  • 2b434c449e whisper : switch back to F32 mask (#0) Georgi Gerganov 2024-05-13 14:43:43 +03:00
  • e93081f83f whisper.android : update example, add field to print timestamp (#2072) zhangjixiong 2024-05-13 19:30:03 +08:00
  • b6bbce4ae9 cmake : fix json INTERFACE library (#2069) Xingchen Song(宋星辰) 2024-05-13 19:29:39 +08:00
  • 7705dc52da main : fix double quote escaping in csv output (#2090) mashizora 2024-05-13 16:55:32 +08:00
  • e6acaf9d91 metal : tune soft_max number of threads (#0) Georgi Gerganov 2024-05-13 11:01:07 +03:00
  • 2c81e6fd51 whisper : remove old flash attn code (#0) Georgi Gerganov 2024-05-13 10:41:33 +03:00
  • 9506267ce5 ggml : try fix ppc64 (#0) Georgi Gerganov 2024-05-12 20:36:31 +03:00
  • fbeb80b5f0 ggml : remove oboslete alibi code (skipme) (#0) Georgi Gerganov 2024-05-12 20:55:57 +03:00
  • 3fa7d29876 talk-llama : sync llama.cpp Georgi Gerganov 2024-05-12 20:12:46 +03:00
  • fe179ae0cc sync : ggml Georgi Gerganov 2024-05-12 19:23:22 +03:00
  • 40aeeeecc4 ggml : optimize for ppc64le using VSX intrinsics (ggml/784) Hong Bo PENG 2024-05-12 17:17:18 +08:00
  • 5a863fbe18 metal : fix indent (ggml/0) Georgi Gerganov 2024-05-11 16:57:53 +03:00
  • 91c646c61d ggml : restore sigmoid decl order (ggml/0) Georgi Gerganov 2024-05-11 16:50:54 +03:00
  • accada542a ggml : resolve merge (ggml/0) Georgi Gerganov 2024-05-11 16:25:50 +03:00
  • e54329da7b ggml : full ALiBi support (llama/7192) Georgi Gerganov 2024-05-11 10:32:41 +03:00
  • 284fac39fb metal : fix flash attention kernel requirements (llama/7169) Georgi Gerganov 2024-05-10 18:20:10 +03:00
  • fe454b8d9e Minor arithmetic improvement to mmvq wrapper kernel (llama/7172) Ouadie EL FAROUKI 2024-05-10 01:32:15 +01:00
  • c114b75aee Vulkan Bugfixes and Improvements (llama/7084) 0cc4m 2024-05-09 20:39:54 +02:00
  • 4be936b88b CUDA: generalize FP16 fattn vec kernel (llama/7061) Johannes Gäßler 2024-05-09 14:32:02 +02:00
  • 26c550f772 opencl : alignment size converted from bits to bytes (llama/7090) Albert Jin 2024-05-09 17:34:37 +08:00
  • 24f0aa460b Introduction of CUDA Graphs to LLama.cpp (llama/6766) agray3 2024-05-08 21:55:49 +01:00
  • 69efc39d5c metal : use vm_allocate instead of posix_memalign on macOS (llama/7078) Gilad S 2024-05-08 22:08:10 +03:00
  • a2ad810118 ggml : introduce bfloat16 support (llama/6412) Justine Tunney 2024-05-08 02:30:09 -04:00
  • 1ae1a9cd56 metal : fix unused warning Georgi Gerganov 2024-05-08 09:14:50 +03:00
  • b5521fea19 Add an option to build without CUDA VMM (llama/7067) William Tambellini 2024-05-06 11:12:14 -07:00
  • 9b84195225 gguf-split: add --no-tensor-first-split (llama/7072) Xuan Son Nguyen 2024-05-04 18:56:22 +02:00
  • 11c1df0436 CUDA: CUDART < 11.7 workaround for __hmax, __hmax2 (llama/7019) Johannes Gäßler 2024-05-01 14:46:37 +02:00
  • c754494fdd switch to using localizedDescription (llama/7010) Kevin Gibbons 2024-04-30 08:14:02 -07:00
  • 1bce67999d metal : remove deprecated error code (llama/7008) Georgi Gerganov 2024-04-30 15:52:21 +03:00
  • 6c39ea46b6 metal : log more info on error (llama/6987) Kevin Gibbons 2024-04-30 02:34:50 -07:00
  • 156a33a990 ggml : add Flash Attention (llama/5021) Georgi Gerganov 2024-04-30 12:16:08 +03:00
  • 5167ebdfca ggml : fix __MSC_VER -> _MSC_VER (llama/6977) Georgi Gerganov 2024-04-29 17:55:02 +03:00
  • b574646d75 Fix more int overflow during quant (PPL/CUDA). (llama/6563) DAN™ 2024-04-28 18:38:44 -04:00
  • 388c3462a6 gguf : enforce that tensor names are unique (llama/6905) Xuan Son Nguyen 2024-04-28 17:36:18 +02:00
  • 9ad202bee9 add device version in device list (llama/6959) Neo Zhang 2024-04-28 22:40:31 +08:00
  • f0d3fb4a7e Reset schedule earlier to allow overlap with ggml graph computation on device (llama/6933) agray3 2024-04-26 19:08:30 +01:00
  • 9d4c8b8aa5 add basic tensor data validation function (llama/6884) slaren 2024-04-26 18:39:58 +02:00
  • ecfac1e240 gguf : fix mismatch between alloc and free functions (llama/6929) slaren 2024-04-26 17:07:42 +02:00
  • 6f7140f568 Merge pull request from GHSA-p5mv-gjc5-mwqv Georgi Gerganov 2024-04-26 10:41:53 +03:00
  • 05b17112cf ggml : fix redefinition of vaddvq_f32 for 32-bit ARM (llama/6906) Georgi Gerganov 2024-04-25 15:48:25 +03:00
  • a15fb5cd79 ggml : fix MIN / MAX macros (llama/6904) Georgi Gerganov 2024-04-25 15:12:28 +03:00
  • 63fd148d8f ggml : move 32-bit arm compat in ggml-impl.h (llama/6865) Georgi Gerganov 2024-04-24 12:00:07 +03:00
  • 6c3971b29b llamafile : improve sgemm.cpp (llama/6796) Justine Tunney 2024-04-22 15:00:36 -04:00
  • a6d264f331 ggml : fix calloc argument ordering. (llama/6820) Dave Airlie 2024-04-23 00:05:06 +10:00
  • 2959686019 ggml : fix ggml_backend_cpu_supports_op() for CPY (llama/0) Georgi Gerganov 2024-04-21 16:47:57 +03:00