Commit Graph

  • 5089ab2d6a whisper : fix build (#0) Georgi Gerganov 2024-11-15 08:40:47 +02:00
  • bdbb906817 sync : ggml Georgi Gerganov 2024-11-15 08:40:34 +02:00
  • fa2ebd336e sycl : Fixes to broken builds and test-backend-ops (llama/10257) Alberto Cabrera Pérez 2024-11-13 09:40:57 +00:00
  • 21b01a21b6 vulkan: Optimize contiguous copies (llama/10254) Jeff Bolz 2024-11-13 00:58:57 -06:00
  • b54ce5edc5 vulkan: Throttle the number of shader compiles during the build step. (llama/10222) Jeff Bolz 2024-11-11 11:13:51 -06:00
  • 26a31b78e9 metal : more precise Q*K in FA vec kernel (llama/10247) Georgi Gerganov 2024-11-11 08:39:13 +02:00
  • 14d13c5f9f vulkan: Fix newly added tests for permuted mul_mat and 1D im2col (llama/10226) Jeff Bolz 2024-11-10 05:37:56 -06:00
  • 5e110c2eb5 metal : reorder write loop in mul mat kernel + style (llama/10231) Georgi Gerganov 2024-11-09 11:53:13 +02:00
  • 4a9926d521 metal : fix build and some more comments (llama/10229) Georgi Gerganov 2024-11-09 11:53:02 +02:00
  • ae3c5642d0 metal : fix F32 accumulation in FA vec kernel (llama/10232) Georgi Gerganov 2024-11-09 11:52:45 +02:00
  • e287a3b627 metal : hide debug messages from normal log Georgi Gerganov 2024-11-09 11:21:49 +02:00
  • b890243690 ggml: fix zero division in ‘dne’ calculation in CUDA COUNT_EQUAL operator when ‘ne’ is small (#10213) SXX 2024-11-09 15:35:46 +08:00
  • b7b38f7d68 ggml : optimize llamafile cpu matrix multiplication for ppc64le (llama/10156) amritahs-ibm 2024-11-09 12:47:50 +05:30
  • 9f67aab211 metal : opt-in compile flag for BF16 (llama/10218) Georgi Gerganov 2024-11-08 21:59:46 +02:00
  • 8f0f785d88 metal : improve clarity (minor) (llama/10171) Georgi Gerganov 2024-11-08 18:37:41 +02:00
  • d0b8335789 metal : optimize FA kernels (llama/10171) Georgi Gerganov 2024-11-08 13:47:22 +02:00
  • 1550be79f1 ggml : add ggml-cpu.h to the public headers (llama/10204) Diego Devesa 2024-11-07 18:16:08 +01:00
  • 807f848c2f fix q4_0_8_8 format for corrupted tokens issue (llama/10198) snadampal 2024-11-07 02:02:08 -06:00
  • 42398f13b0 Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleration (llama/10133) Zhiyuan Li 2024-11-07 18:19:10 +11:00
  • 31c3482a4e metal : add BF16 support (llama/8439) Georgi Gerganov 2024-11-06 19:53:51 +02:00
  • 50257af686 metal : fix from ptr buffer name (llama/10189) Diego Devesa 2024-11-06 12:10:07 +01:00
  • d111a0987e ggml : adjust is_first_call init value (llama/10193) Georgi Gerganov 2024-11-06 11:20:10 +02:00
  • 915bcd2c63 metal : add quantized FA support (llama/10149) Georgi Gerganov 2024-11-06 10:24:23 +02:00
  • f69c8b6f1b ggml : fix arch check in bf16_to_fp32 (llama/10164) Diego Devesa 2024-11-04 23:17:01 +01:00
  • 8c9044bef0 Q6_K AVX improvements (llama/10118) Eve 2024-11-04 22:06:31 +00:00
  • 5f8e928194 ggml : fix gelu tables initialization (llama/10172) Diego Devesa 2024-11-04 20:06:58 +01:00
  • 25da30bd60 ggml : fix q4xx mat mul, increase ggml_aligned_malloc alignment (llama/10167) Diego Devesa 2024-11-04 17:34:08 +01:00
  • 542734100e fix build break on arm64 linux (llama/10166) snadampal 2024-11-04 09:08:33 -06:00
  • b06b4c0c08 cuda : clear error after changing peer access (llama/10153) Diego Devesa 2024-11-04 13:10:23 +01:00
  • 939d36fb4c metal : simplify f16 and f32 dequant kernels (llama/0) Georgi Gerganov 2024-11-04 13:49:34 +02:00
  • 1471e41180 metal : move dequantize templates to beginning of MSL source (llama/0) Georgi Gerganov 2024-11-04 13:43:32 +02:00
  • 35949192e9 CANN: adjust backend registry refactor. (llama/10158) leo-pony 2024-11-04 19:08:22 +08:00
  • 9c817edb48 ggml : move CPU backend to a separate file (llama/10144) Diego Devesa 2024-11-03 19:34:08 +01:00
  • 24a0feb5d9 metal : minor fixup in FA kernel (llama/10143) Georgi Gerganov 2024-11-03 15:18:40 +02:00
  • 2ab8cce7e3 llama : add simple-chat example (llama/10124) Diego Devesa 2024-11-01 23:50:59 +01:00
  • b40c255e98 llama : use smart pointers for ggml resources (llama/10117) Diego Devesa 2024-11-01 23:48:26 +01:00
  • ec3e16445e vulkan : improve ggml_vk_create_buffer error handling (llama/9898) Shupei Fan 2024-11-02 02:33:14 +08:00
  • 0665168ef3 ggml : remove ggml_scratch (llama/10121) Georgi Gerganov 2024-11-01 12:58:45 +02:00
  • 5f6b992eea build: fix build error in Windows env with OneAPI setup (llama/10107) Zhenwei Jin 2024-11-01 11:09:59 +08:00
  • 3e231ab9cc llama : fix buffer checks for mamba and rwk (llama/10111) Diego Devesa 2024-10-31 22:54:23 +01:00
  • 371bfaca8c ggml : check tensor name lengths in gguf files (llama/10100) Diego Devesa 2024-10-31 11:40:59 +01:00
  • 91e30a3a23 kompute: add mul_mat_q4_k shader (llama/10097) Sergio López 2024-10-31 10:09:52 +01:00
  • 1e122d66f9 kompute: add backend registry / device interfaces (llama/10045) Sergio López 2024-10-30 17:01:52 +01:00
  • 63a4e09a0f ggml : fix memory leaks when loading invalid gguf files (llama/10094) Diego Devesa 2024-10-30 14:51:21 +01:00
  • 75dd198870 ggml : add Q4_0_8_8 RISC-V GEMV and GEMM kernels (llama/10029) xctan 2024-10-30 15:00:40 +08:00
  • 1d48457aa6 llama : refactor model loader with backend registry (llama/10026) Diego Devesa 2024-10-30 02:01:23 +01:00
  • 307712a903 ggml: Add POOL2D OP for GPU acceleration to the Vulkan backend in the MobileVLM model. (llama/9763) Changyeon Kim 2024-10-29 17:52:56 +09:00
  • fbc9a05ddf musa: workaround for Guilty Lockup in cleaning src0 (llama/10042) R0CKSTAR 2024-10-28 17:02:48 +08:00
  • 28496ac55e cmake : make it possible linking ggml as external lib (ggml/1003) Yuri Khrustalev 2024-11-02 05:09:12 -04:00
  • b1c06c09b0 metal : fix minor string leaks (ggml/1004) Plamen Minev 2024-11-01 16:55:10 +02:00
  • 498ac0dc27 scripts : update sync Georgi Gerganov 2024-11-15 08:34:49 +02:00
  • 03af461de8 ci : fix building workflow for linux/arm64 container (#2555) Raiya Araki 2024-11-15 18:07:17 +09:00
  • f19463ece2 ruby : extend API (#2551) KITAITI Makoto 2024-11-14 04:52:56 +09:00
  • 5f8a086e22 whisper.swiftui : add model download list & bench methods (#2546) Jhen-Jie Hong 2024-11-14 03:51:34 +08:00
  • a28d82e373 ruby : fix the instructions (#2548) Wilson Silva 2024-11-13 19:47:42 +00:00
  • 5ccca19f0c ggml : vulkan logs (#2547) thewh1teagle 2024-11-13 21:47:15 +02:00
  • 300c07b94d examples : fix ffmpeg v5 build (#2543) Stefan Sydow 2024-11-13 20:41:52 +01:00
  • 31aea563a8 whisper : fix extra memory usage (#2534) Vin Misra 2024-11-06 13:02:11 -08:00
  • 0377596b77 whisper : backend registry init before model load Georgi Gerganov 2024-10-31 22:53:46 +02:00
  • c65d0fd3c8 talk-llama : sync llama.cpp Georgi Gerganov 2024-10-31 22:29:22 +02:00
  • d9efb664ac sync : ggml Georgi Gerganov 2024-10-31 22:26:28 +02:00
  • b5b4b0f5de ggml : add AMX backend (llama/8998) Ma Mingfei 2024-10-26 09:43:40 +03:00
  • ab36d02560 metal : support permuted matrix multiplicaions (llama/10033) Georgi Gerganov 2024-10-25 22:26:15 +03:00
  • 6e67749c00 CUDA: fix insufficient buffer clearing for MMQ (llama/10032) Johannes Gäßler 2024-10-24 14:40:23 +02:00
  • ab0385f43b CUDA: fix MMQ for non-contiguous src0, add tests (llama/10021) Johannes Gäßler 2024-10-24 11:09:36 +02:00
  • 10eb603a3c increase cuda_cpy block size (ggml/996) bssrdf 2024-10-23 14:34:00 -04:00
  • a3231b2f2e metal : add POOL2D and fix IM2COL (llama/9943) Jun Hee Yoo 2024-10-23 19:33:45 +09:00
  • 13db492f83 Adapt to dynamically loadable backends mechanism (llama/9970) leo-pony 2024-10-22 16:16:01 +08:00
  • 741c138aa1 ggml : add asserts for type conversion in fattn kernels (llama/9971) Georgi Gerganov 2024-10-21 16:20:46 +03:00
  • 25f9fee6fb rpc : pack only RPC structs (llama/9959) Radoslav Gerganov 2024-10-21 13:35:40 +03:00
  • 7c1570bee6 fix mul_mat_vec_q and *_vec_q error (llama/9939) Neo Zhang Jianyu 2024-10-21 14:26:09 +08:00
  • 4078e4c388 rpc : backend refactoring (llama/9912) Radoslav Gerganov 2024-10-18 14:33:58 +03:00
  • a4a22daa8f Add SYCL Backend registry, device and Event Interfaces (llama/9705) Ouadie EL FAROUKI 2024-10-18 06:46:16 +01:00
  • e1936eb2a5 add amx kernel for gemm (llama/8998) Ma Mingfei 2024-10-18 13:34:36 +08:00
  • 28b044dad9 vulkan : add backend registry / device interfaces (llama/9721) Diego Devesa 2024-10-17 02:46:58 +02:00
  • b8f11a0a17 fix: allocating CPU buffer with size 0 (llama/9917) Gilad S 2024-10-17 02:34:22 +03:00
  • ff5a838099 fix: use vm_allocate to allocate CPU backend buffer on macOS (llama/9875) Gilad S 2024-10-17 01:36:51 +03:00
  • 84713613be CUDA: fix 1D im2col, add tests (ggml/993) Johannes Gäßler 2024-10-18 09:24:44 +02:00
  • ded89c9d08 Fix cann compilation error (llama/9891) leo-pony 2024-10-16 08:51:46 +08:00
  • 042e95d92f Vectorize load instructions in dmmv f16 CUDA kernel (llama/9816) agray3 2024-10-14 01:49:08 +01:00
  • 81110c0174 ggml : move more prints to the ggml log system (llama/9839) Diego Devesa 2024-10-11 15:34:45 +02:00
  • c313723860 rpc : add backend registry / device interfaces (llama/9812) Diego Devesa 2024-10-10 20:14:55 +02:00
  • e69b2371e2 musa: add docker image support (llama/9685) R0CKSTAR 2024-10-11 02:10:37 +08:00
  • 1531259b2c ggml : fix BLAS with unsupported types (llama/9775) Diego Devesa 2024-10-08 14:21:43 +02:00
  • 44bc2767fd ggml : add backend registry / device interfaces to BLAS backend (llama/9752) Diego Devesa 2024-10-07 21:55:08 +02:00
  • bd7ace7adc Update building for Android (llama/9672) Andrew Minh Nguyen 2024-10-07 09:37:31 -07:00
  • 315364d7de ggml : add metal backend registry / device (llama/9713) Georgi Gerganov 2024-10-07 18:27:51 +03:00
  • 80753d4da8 metal : single allocation of encode_async block (llama/9747) Paul Tsochantaris 2024-10-07 13:26:31 +01:00
  • 8f9bdca4c4 ggml-alloc : remove buffer_id from leaf_alloc (ggml/987) Daniel Bevenius 2024-10-09 16:40:35 +02:00
  • 4e10afb5a9 scripts : sync amx Georgi Gerganov 2024-10-31 22:13:24 +02:00
  • aa037a60f3 ggml : alloc ggml_contexts on the heap (#2525) Georgi Gerganov 2024-10-31 22:00:09 +02:00
  • 552419f2c0 ggml : aligned malloc -> malloc gg/reduce-ctx-use Georgi Gerganov 2024-10-31 21:40:11 +02:00
  • 987f3145d0 ggml : allocate contexts on the heap (v2) Georgi Gerganov 2024-10-31 12:46:20 +02:00
  • 3689d49b81 whisper : reduce ggml_context usage Georgi Gerganov 2024-10-30 13:39:14 +02:00
  • 19dca2bb14 ci : fix openblas build (#2511) Georgi Gerganov 2024-10-30 12:58:26 +02:00
  • 55e422109b scripts : add turbo-q8_0 to the benchmark Georgi Gerganov 2024-10-29 19:37:24 +02:00
  • 3f020fac9d whisper : minor compile warning Georgi Gerganov 2024-10-29 19:27:52 +02:00
  • 1626b73b03 whisper : move new-segment callback after DTW step (#2515) jettoblack 2024-10-29 02:47:21 -04:00
  • 850f7b19d3 ruby : fix installation test (#2519) KITAITI Makoto 2024-10-29 15:45:37 +09:00
  • d4bc413505 ruby : add more APIs (#2518) KITAITI Makoto 2024-10-29 02:23:23 +09:00