upstream-archive/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-07-29 14:11:13 -05:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • adffa6ffd5 common : improve -ctv -ctk CLI arguments (#10806) b4317 Xuan Son Nguyen 2024-12-12 22:53:05 +01:00
  • 274ec65af6 contrib : add ngxson as codeowner (#10804) Xuan Son Nguyen 2024-12-12 20:52:28 +01:00
  • 8faa1d4dd4 CUDA: faster non-contiguous concat (#10760) b4315 a3sh 2024-12-13 02:09:50 +08:00
  • cb13ef85a4 remove CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS (#10797) b4314 Diego Devesa 2024-12-12 19:02:49 +01:00
  • 4064c0e3b6 Vulkan: Use improved q4_k and q5_k dequant code in dequant shaders (#10798) 0cc4m 2024-12-12 18:36:00 +01:00
  • dc5301d565 Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats (#10721) b4312 0cc4m 2024-12-12 18:35:37 +01:00
  • 9fdb124304 common : add missing env var for speculative (#10801) b4311 Xuan Son Nguyen 2024-12-12 16:57:32 +01:00
  • 5555c0c1f6 docs: update server streaming mode documentation (#9519) CentricStorm 2024-12-11 22:40:40 +00:00
  • 973f328b1e Merge pull request #10788 from ggerganov/gg/gguf-py-0.11.0 Georgi Gerganov 2024-12-11 23:14:46 +02:00
  • fb18934a97 gguf-py : bump version to 0.11.0 gguf-v0.11.0 gg/gguf-py-0.11.0 Georgi Gerganov 2024-12-11 23:13:31 +02:00
  • 235f6e14bf server : (UI) add tok/s, get rid of completion.js (#10786) gguf-py gguf ggu Xuan Son Nguyen 2024-12-11 20:52:14 +01:00
  • 1a31d0dc00 Update README.md (#10772) qingy1337 2024-12-11 07:16:32 -08:00
  • 92f77a640f ci : pin nodejs to 22.11.0 (#10779) Xuan Son Nguyen 2024-12-11 14:59:41 +01:00
  • 484d2f31ae bug-fix: snprintf prints NULL in place of the last character (#10419) b4304 kallewoof 2024-12-11 22:48:04 +09:00
  • 4b4d92b098 docs: fix server documentation formatting (#10776) CentricStorm 2024-12-11 10:47:43 +00:00
  • 43041d2eb3 ggml: load all backends from a user-provided search path (#10699) b4302 Gilad S. 2024-12-11 02:47:21 +02:00
  • 4f3a7e279b Force max subgroup size for coopmat shaders 0cc4m/vulkan-subgroup-size-control-amd 0cc4m 2024-12-10 20:27:04 +00:00
  • b685daf386 vulkan: request round-to-even for fp16 in im2col/rope_head (#10767) b4301 Jeff Bolz 2024-12-10 14:23:17 -06:00
  • 2dc175fb2b Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats 0cc4m 2024-12-08 14:41:40 +00:00
  • dafae66cc2 vulkan: dynamic subgroup size for the remaining k quants (#10745) b4300 Eve 2024-12-10 19:33:23 +00:00
  • ae4b922614 imatrix : Add imatrix to --no-context-shift (#10766) b4299 Bartowski 2024-12-10 12:23:50 -05:00
  • 750cb3e246 CUDA: rename macros to avoid conflicts with WinAPI (#10736) b4298 Andreas Kieslinger 2024-12-10 18:23:24 +01:00
  • a86ad841f1 server : add flag to disable the web-ui (#10762) (#10751) b4297 Yüg 2024-12-10 17:22:34 +00:00
  • a05e2afcc2 vulkan: disable spirv-opt for coopmat shaders (#10763) b4296 Jeff Bolz 2024-12-10 11:22:20 -06:00
  • 26a8406ba9 CUDA: fix shared memory access condition for mmv (#10740) b4295 Johannes Gäßler 2024-12-09 20:07:12 +01:00
  • c37fb4cf62 Changes to CMakePresets.json to add ninja clang target on windows (#10668) Srihari-mcw 2024-12-09 23:10:19 +05:30
  • 3d98b4cb22 vulkan: fix compile warnings (#10731) b4293 Jeff Bolz 2024-12-09 01:24:01 -06:00
  • 1a05004743 cmake : simplify msvc charsets (#10672) b4292 Borislav Stanimirov 2024-12-09 09:15:13 +02:00
  • ce8784bdb1 server : fix format_infill (#10724) b4291 Xuan Son Nguyen 2024-12-08 23:04:29 +01:00
  • b8d1b1a5e1 server : fix infill prompt format gg/server-fix-infill Georgi Gerganov 2024-12-08 22:12:11 +02:00
  • e52522b869 server : bring back info of final chunk in stream mode (#10722) b4290 Xuan Son Nguyen 2024-12-08 20:38:51 +01:00
  • 06d70147e6 Vulkan: fix NaN in tanh.comp with AMD proprietary driver on Windows (#10723) stduhpf 2024-12-08 19:19:19 +01:00
  • 43ed389a3f llama : use cmake for swift build (#10525) b4288 Diego Devesa 2024-12-08 12:14:54 +01:00
  • ecc93d0558 vulkan: compile a test shader in cmake to check for coopmat2 support (#10713) b4287 Jeff Bolz 2024-12-08 02:05:55 -06:00
  • a6648b9df7 server : chunked prefill support gg/server-chunked-prefill Georgi Gerganov 2024-12-08 09:48:18 +02:00
  • 62e84d9848 llama : add 128k yarn context for Qwen (#10698) Robert Collins 2024-12-07 16:12:27 -05:00
  • 3573fa8e7b server : (refactor) no more json in server_task input (#10691) b4285 Xuan Son Nguyen 2024-12-07 20:21:09 +01:00
  • d9c3ba2b77 ggml : disable iq4_nl interleave size 8 (#10709) b4284 Georgi Gerganov 2024-12-07 18:38:15 +02:00
  • ce4a7b8493 server : various fixes (#10704) b4283 Georgi Gerganov 2024-12-07 18:02:05 +02:00
  • 19d8762ab6 ggml : refactor online repacking (#10446) b4282 Djip007 2024-12-07 13:37:50 +01:00
  • c2a16c0bdb server : fix free of spec context and batch (#10651) b4281 Georgi Gerganov 2024-12-07 11:52:44 +02:00
  • 3df784b305 Vulkan: VK_KHR_cooperative_matrix support to speed up prompt processing (#10597) b4280 0cc4m 2024-12-07 10:24:15 +01:00
  • 86a1934978 metal : Extend how Llama.cpp locates metal resources (#10676) b4279 Robert Ormandi 2024-12-07 01:55:01 -06:00
  • 784a14aa49 convert : add support for Roberta embeddings (#10695) Sukriti Sharma 2024-12-07 00:02:14 -07:00
  • c5ede3849f convert : add custom attention mapping Georgi Gerganov 2024-12-06 21:33:15 +02:00
  • f162d45a21 common : bring back --no-warmup to server (#10686) b4276 Xuan Son Nguyen 2024-12-06 13:29:05 +01:00
  • 6c5bc0625f server : (refactoring) do not rely on JSON internally (#10643) Xuan Son Nguyen 2024-12-06 11:14:32 +01:00
  • 7736837d62 fix(server) : not show alert when DONE is received (#10674) Plamen Minev 2024-12-05 23:36:41 +02:00
  • c9c6e01dae vulkan: Add VK_NV_cooperative_matrix2 support for mul_mat and flash attention (#10206) b4273 Jeff Bolz 2024-12-05 13:15:05 -06:00
  • 6fe6247831 llama : add Minerva 7B model support (#10673) b4272 Riccardo Orlando 2024-12-05 19:30:59 +01:00
  • 0cd182ebcc sync : ggml b4271 Georgi Gerganov 2024-12-05 13:27:42 +02:00
  • a8cbab201d ggml: add GGML_SET Metal kernel + i32 CPU kernel (ggml/1037) PAB 2024-12-04 09:19:30 +01:00
  • c2082d93a8 ggml : add GGML_PAD_REFLECT_1D operation (ggml/1034) PAB 2024-12-03 20:20:04 +01:00
  • d405804be8 py : update outdated copy-paste instructions [no ci] (#10667) Daniel Bevenius 2024-12-05 08:47:55 +01:00
  • f112d198cd Update deprecation-warning.cpp (#10619) b4267 aryantandon01 2024-12-05 03:49:20 +05:30
  • 1da7b76569 server : fix speculative decoding with context shift (#10641) b4266 Georgi Gerganov 2024-12-04 22:38:20 +02:00
  • a8046c888a use calloc instead of malloc jg/gguf-refactor Johannes Gäßler 2024-12-04 17:24:35 +01:00
  • 59f4db1088 ggml : add predefined list of CPU backend variants to build (#10626) b4265 Diego Devesa 2024-12-04 14:45:40 +01:00
  • 2803540814 ggml-cpu : fix HWCAP2_I8MM value (#10646) Diego Devesa 2024-12-04 14:40:44 +01:00
  • 096b847a0f fix wrong type in print Johannes Gäßler 2024-12-04 14:16:05 +01:00
  • b88727009d GGUF: backend support, fixed-width I/O, misc fixes Johannes Gäßler 2024-12-03 21:43:57 +01:00
  • 81611bef72 server : add tests gg/server-fix-spec-ctx-shift Georgi Gerganov 2024-12-04 13:11:26 +02:00
  • 253b7fde91 Fix HF repo commit to clone lora test models (#10649) ltoniazzi 2024-12-04 09:45:48 +00:00
  • 8d0cfd554a llama: Support MiniCPM-1B (with & w/o longrope) (#10559) b4262 JFLFY2255 2024-12-04 17:42:50 +08:00
  • b436edaad9 server : take into account speculative limits Georgi Gerganov 2024-12-04 10:44:48 +02:00
  • 2759916d86 vulkan: Implement "fast divide" (mul+shift) for unary ops like copy (#10642) b4261 Jeff Bolz 2024-12-04 01:28:59 -06:00
  • 40c6d79fb5 SYCL : Move to compile time oneMKL interface backend selection for NVIDIA backend (#10584) b4260 Nicolò Scipione 2024-12-04 02:29:20 +01:00
  • 98036d5670 fix typo of README.md (#10605) Wang Ran (汪然) 2024-12-04 09:22:50 +08:00
  • cd2f37b304 Avoid using __fp16 on ARM with old nvcc (#10616) b4258 Frankie Robertson 2024-12-04 02:41:37 +02:00
  • da6aac91f1 Add docs for creating a static build (#10268) (#10630) Benson Wong 2024-12-03 16:40:36 -08:00
  • 01e6d9bb71 clip : add sycl support (#10574) b4256 piDack 2024-12-04 08:26:37 +08:00
  • a5a915b51e server : fix speculative decoding with context shift Georgi Gerganov 2024-12-03 22:44:19 +02:00
  • cc98896db8 vulkan: optimize and reenable split_k (#10637) b4255 Jeff Bolz 2024-12-03 13:29:54 -06:00
  • 91c36c269b server : (web ui) Various improvements, now use vite as bundler (#10599) b4254 Xuan Son Nguyen 2024-12-03 19:38:44 +01:00
  • 1cd3df46bd scripts : remove amx sync b4253 Georgi Gerganov 2024-12-03 19:42:30 +02:00
  • c505471857 sync : ggml Georgi Gerganov 2024-12-03 19:40:25 +02:00
  • e9e661bd59 CUDA: remove unnecessary warp reduce in FA (ggml/1032) mahorozte 2024-12-03 21:11:43 +08:00
  • efb6ae9630 feat: add GGML_UNARY_OP_ARGMAX Metal kernel (ggml/1019) PAB 2024-12-02 19:27:24 +01:00
  • 667d70d170 metal : add GGML_OP_CONV_TRANSPOSE_1D kernels (ggml/1026) PAB 2024-11-28 09:25:06 +01:00
  • 3b4f2e33e2 llama : add missing LLAMA_API for llama_chat_builtin_templates (#10636) b4248 Xuan Son Nguyen 2024-12-03 12:54:30 +01:00
  • 82bca2257b readme : add option, update default value, fix formatting (#10271) Nikolaos Pothitos 2024-12-03 12:50:08 +02:00
  • 0115df2f65 metal : small-batch mat-mul kernels (#10581) b4246 Georgi Gerganov 2024-12-03 11:52:33 +02:00
  • 515d4e5372 github : minify link [no ci] (revert) Georgi Gerganov 2024-12-03 11:21:43 +02:00
  • 844e2e1fee github : minify link [no ci] Georgi Gerganov 2024-12-03 11:20:35 +02:00
  • 70b98fadbc server : fix default draft model parameters (#10586) b4243 Georgi Gerganov 2024-12-03 11:20:00 +02:00
  • 33d7b70c88 server : do not speculate during prompt processing gg/server-fix-spec Georgi Gerganov 2024-12-03 10:58:43 +02:00
  • 642330ac7c llama : add enum for built-in chat templates (#10623) b4242 Xuan Son Nguyen 2024-12-02 22:10:19 +01:00
  • 8648c52101 make : deprecate (#10514) Georgi Gerganov 2024-12-02 21:22:53 +02:00
  • 64ed2091b2 server: Add "tokens per second" information in the backend (#10548) b4240 haopeng 2024-12-02 21:45:54 +08:00
  • 991f8aabee SYCL: Fix and switch to GGML_LOG system instead of fprintf (#10579) b4239 Akarshan Biswas 2024-12-02 12:34:11 +05:30
  • 4cb003dd8d contrib : refresh (#10593) Georgi Gerganov 2024-12-02 08:53:27 +02:00
  • 917786f43d Add mistral-v1, mistral-v3, mistral-v3-tekken and mistral-v7 chat template types (#10572) Juk Armstrong 2024-12-01 22:09:49 +00:00
  • 5e1ed95583 grammars : add English-only grammar (#10612) Georgi Gerganov 2024-12-01 21:37:54 +02:00
  • 5c7a5aa0c3 ci: add error handling for Python venv creation in run.sh (#10608) Wang Qin 2024-12-01 10:11:42 -08:00
  • 3420909dff ggml : automatic selection of best CPU backend (#10606) b4234 Diego Devesa 2024-12-01 16:12:41 +01:00
  • 86dc11c5bc server : bind to any port when specified (#10590) b4233 alek3y 2024-12-01 12:33:12 +01:00
  • 6acce39710 readme : update the usage section with examples (#10596) Georgi Gerganov 2024-12-01 11:25:17 +02:00
  • 43957ef203 build: update Makefile comments for C++ version change (#10598) b4231 Wang Qin 2024-11-30 19:19:44 -08:00
  • 0c39f44d70 ggml-cpu: replace AArch64 NEON assembly with intrinsics in ggml_gemv_q4_0_4x4_q8_0() (#10567) b4230 Adrien Gallouët 2024-11-30 18:13:18 +01:00
  • 3e0ba0e604 readme : remove old badge Georgi Gerganov 2024-11-30 10:09:21 +02:00