vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Commit graph

  • b1c66e4ad7 Merge branch 'master' into concedo_experimental Concedo 2023-09-03 11:47:49 +08:00
  • b61ae67082 use HIP visible devices, updated lite Concedo 2023-09-03 11:46:19 +08:00
  • 8eaa401def feat: embed ggml-metal.metal Meng Zhang 2023-09-03 10:41:03 +08:00
  • 24215ae81d Add missing c file to Package.swift kchro3 2023-09-02 18:32:11 -07:00
  • c031b6c5d9 compatibility fixes for MAP_ANONYMOUS slaren 2023-09-02 23:42:54 +02:00
  • 96f3662632 ggml-alloc : use virtual memory for measurement slaren 2023-09-02 22:28:36 +02:00
  • 80ac697df9
         move measurement memory segment to upper region of the address space xaedes 2023-09-02 21:44:20 +02:00
  • 2d2bdc0df7
         remove unnecessary "0x" before "%p" output xaedes 2023-09-02 21:28:08 +02:00
  • 1ce7023eed
         revert last commit xaedes 2023-09-02 21:27:12 +02:00
  • 8d982c8fd9
         bug fix, probably solves the 'ggml_allocr_alloc: not enough space in the buffer' issue xaedes 2023-09-02 20:53:14 +02:00
  • ded6382961
         add some more allocator debug prints xaedes 2023-09-02 20:52:25 +02:00
  • 3358c381f6
         logging: Fix creating empty file even when disabled (#2966) b1154 Kerfuffle 2023-09-02 11:53:55 -06:00
  • 142d79b459 Merge branch 'master' of https://github.com/ggerganov/llama.cpp into ap-fim apaz-cli 2023-09-02 12:24:22 -05:00
  • 9c0c4b0899
         Minor formatting fix Kerfuffle 2023-09-02 11:11:39 -06:00
  • 363f0bf558 Massive improvement for TG for fp16 Iwan Kawrakow 2023-09-02 18:14:41 +03:00
  • cfe217f1ca
         fix README.md xaedes 2023-09-02 16:11:31 +02:00
  • 6ee12b158b
         increase measured alloc size by tensor_alignment xaedes 2023-09-02 15:59:14 +02:00
  • 52315a4216
         readme : update clblast instructions (#2903) bandoti 2023-09-02 09:53:18 -03:00
  • 8b56b4f2c3
         metal : show all Metal device instances in the system (#2952) Karsten Weiss 2023-09-02 14:29:09 +02:00
  • 7930818a0c
         Update ggml-metal.m Georgi Gerganov 2023-09-02 15:25:40 +03:00
  • 21f3d1be86
         k-quants : fix build on armv7 (android only) (#2920) b1151 Jhen-Jie Hong 2023-09-02 20:23:45 +08:00
  • 7aa864f2d7 docs: add catai to README.md Ido S 2023-09-02 15:05:48 +03:00
  • 019f30dfcb logging: Fix creating empty file even when disabled KerfuffleV2 2023-09-02 03:18:36 -06:00
  • 01eed465c4
         Merge branch 'master' into ik/more_metal_optimizations Georgi Gerganov 2023-09-02 11:22:21 +03:00
  • e5fe4974bb ggml-alloc : use 4g for MEASURE_MAX_SIZE in 32-bit arm Jhen 2023-09-02 11:30:27 +08:00
  • eed651494e Merge branch 'master' into concedo_experimental Concedo 2023-09-02 11:24:28 +08:00
  • 8df03ed026 tweaks for rocm blas Concedo 2023-09-02 09:22:32 +08:00
  • 6f5dc9bc8d k-quants : avoid some unused vzero / mzero define Jhen 2023-09-02 09:20:16 +08:00
  • 571083f508
         server : avoid aniprompt in probabilities of final response (#2849) b1150 Jhen-Jie Hong 2023-09-02 08:31:46 +08:00
  • 9c6ea3b52c Merge branch 'master' into armv7-qk Jhen 2023-09-02 08:24:20 +08:00
  • f04d002844
         cuda : vsubss4 for older versions of ROCm/clang (#2942) b1149 Engininja2 2023-09-01 15:33:19 -06:00
  • ed3fe45f95 make : fix inverted conditional Cebtenzzre 2023-08-30 12:19:55 -04:00
  • b396783a65 make : support overriding CFLAGS/CXXFLAGS/CPPFLAGS/LDFLAGS Cebtenzzre 2023-08-29 16:54:32 -04:00
  • bc6bcdf75a make : pass Raspberry Pi arch flags to g++ as well Cebtenzzre 2023-08-29 16:37:44 -04:00
  • 8fab099ab7 make : put preprocessor stuff in CPPFLAGS Cebtenzzre 2023-08-29 13:04:01 -04:00
  • dca3c06855 make : remove unused -DGGML_BIG_ENDIAN Cebtenzzre 2023-08-29 12:59:06 -04:00
  • 6d06695c7e initializer syntax Concedo 2023-09-02 00:41:50 +08:00
  • 8ee7232c86 Merge remote-tracking branch 'origin' into cmake-find-package Mason M 2023-09-01 13:14:28 -03:00
  • c0b25c6d05 Update README Mason M 2023-09-01 13:09:40 -03:00
  • b520aa7443 Update README Mason M 2023-09-01 13:02:53 -03:00
  • 443b93f48e Add example project using CMake package Mason M 2023-09-01 12:28:17 -03:00
  • 74df0de9e6 Minor Iwan Kawrakow 2023-09-01 18:15:45 +03:00
  • c32ad44f84
         print time per iteration and estimate remaining time xaedes 2023-09-01 17:03:36 +02:00
  • b557bc326d Another attempt Iwan Kawrakow 2023-09-01 17:50:21 +03:00
  • 2b601702a8 Quite significant PP speedup on metal Iwan Kawrakow 2023-09-01 17:06:53 +03:00
  • e3ff8c20c8 Another very minor speedup on metal Iwan Kawrakow 2023-09-01 14:54:10 +03:00
  • 2cb47e0e16 Very minor speedup via simd-group synchronization in f16 x f32 Iwan Kawrakow 2023-09-01 13:21:50 +03:00
  • bcf62ba7b4
         make : try to fix build on Linux Georgi Gerganov 2023-09-01 17:42:32 +03:00
  • e966ae0574
         build : on Mac OS enable Metal by default Georgi Gerganov 2023-08-30 13:11:42 +03:00
  • 6809eb7de9
         Merge branch 'master' into finetune-lora xaedes 2023-09-01 16:07:05 +02:00
  • 69fdbb9abc
         readme : quick start command fix (#2908) ZHAOKAI WANG 2023-09-01 22:06:44 +08:00
  • 7acb1241c6
         update README.md xaedes 2023-09-01 16:04:08 +02:00
  • 5d6f19f16b
         Allow quantize to only copy tensors, some other improvements (#2931) b1147 Kerfuffle 2023-09-01 08:02:48 -06:00
  • 6cbf55a64b
         add finetune to Makefile xaedes 2023-09-01 16:02:45 +02:00
  • 0d58936686
         llama2c : rename function b1146 Georgi Gerganov 2023-09-01 17:00:40 +03:00
  • 5bba329e58
         finetune: automatically allocate all memory and changes to command line options xaedes 2023-09-01 15:58:24 +02:00
  • 1e05731a33 quantize: Use stdout for help message. KerfuffleV2 2023-09-01 07:57:41 -06:00
  • 6c9c23429b
         make : use unaligned vector moves on MinGW (#2945) b1145 Cebtenzzre 2023-09-01 09:53:14 -04:00
  • ee8654bcd0
         minor : add const qualifiers (#2853) b1144 m3ndax 2023-09-01 15:47:27 +02:00
  • 9ec17d8514
         llama2c : remove misleading const Georgi Gerganov 2023-09-01 16:46:30 +03:00
  • 7e01d11a28
         add ggml-alloc API function 'ggml_allocr_max_size' to get max size of alloc xaedes 2023-09-01 15:42:40 +02:00
  • d554a70f11
         initialize opt ggml context if none was provided xaedes 2023-09-01 15:41:57 +02:00
  • 49bb9cbe0f
         docs : add java-llama.cpp to README.md (#2935) Konstantin Herud 2023-09-01 15:36:14 +02:00
  • ef15649972
         build : fix most gcc and clang warnings (#2861) b1142 Cebtenzzre 2023-09-01 09:34:50 -04:00
  • d8d6977f48
         examples : add C grammar (#2357) Ben Siraphob 2023-09-01 09:32:14 -04:00
  • 5aec2cfaac
         ggml : add RISC-V vector intrinsics support (#2929) b1140 Tameem 2023-09-01 18:27:40 +05:00
  • 9ee87b4cee
         Changed assignment clobbering LDFLAGS JohnnyB 2023-09-01 13:29:58 +01:00
  • b7571fde64 Remove trailing whitespace Mason M 2023-09-01 09:06:05 -03:00
  • b46ae7bde9
         Merge branch 'master' into HEAD Georgi Gerganov 2023-09-01 13:43:09 +03:00
  • 13268c5331
         metal : slight speed-up for add and mul kernels (#2917) Georgi Gerganov 2023-09-01 13:42:41 +03:00
  • 048831e95b
         Merge branch 'master' into HEAD Georgi Gerganov 2023-09-01 13:31:09 +03:00
  • a3880cab25
         Merge 8c2b881281 into 4dcd47d71d Georgi Gerganov 2023-09-01 12:23:04 +02:00
  • 7bb41cf099 [ggml-openc.cpp]: fix a bug in ggml_cl_pool_malloc() for ggml_cl_mul_mat_f32() Wentai Zhang 2023-09-01 18:10:56 +08:00
  • 339d3477bf ggml_metal_init: Show all Metal device instances in the system Karsten Weiss 2023-09-01 10:59:13 +02:00
  • 4dcd47d71d
         logs : fix mingw-like builds (fixes #2898) (#2911) b1138 staviq 2023-09-01 11:07:06 +02:00
  • 18705a30ef
         llama2c : fix segfault and alloc-dealloc-mismatch (#2913) b1137 Cebtenzzre 2023-09-01 05:03:49 -04:00
  • 5925c23d51 fix for RWKV Concedo 2023-09-01 17:02:11 +08:00
  • e8d9158925
         metal: somewhat faster f16 x f32 matrix multiply kernel (#2951) Kawrakow 2023-09-01 11:15:57 +03:00
  • cad50d1971 Better use 32 thread groups for f16 x f32 Iwan Kawrakow 2023-09-01 10:44:13 +03:00
  • 81abd3cb1f Merge remote-tracking branch 'elbios/concat_output_mutex' into concedo_experimental Concedo 2023-09-01 15:24:13 +08:00
  • d7fed4732f fix for typical sampler Concedo 2023-09-01 15:24:00 +08:00
  • af226bd26e Somewhat faster f16 x f32 matrix multiply kernel Iwan Kawrakow 2023-09-01 10:14:42 +03:00
  • 759dcdcfdb
         fix typo JackJollimore 2023-09-01 03:19:39 -03:00
  • b47ee69572
         Ensure device compatibility for CLBlast JackJollimore 2023-09-01 02:19:47 -03:00
  • 30588617fb Fix race condition by locking concat_output string Elbios 2023-09-01 07:08:50 +02:00
  • 736ff7c103
         stylized OpenBlas text JackJollimore 2023-09-01 00:32:40 -03:00
  • d719f6cd12
         easily copy/paste export JackJollimore 2023-09-01 00:23:28 -03:00
  • bce1fef328
         convert : fix another python 3.8 issue (#2949) Cebtenzzre 2023-08-31 22:13:51 -04:00
  • 7ee8d35834 convert : fix another python 3.8 issue Cebtenzzre 2023-08-31 22:02:22 -04:00
  • 673f424f34
         typo fix JackJollimore 2023-08-31 22:18:26 -03:00
  • 5a3dc43bfe
         Clean/Update Android ClBlast section JackJollimore 2023-08-31 22:13:31 -03:00
  • 378e9bce73 quick start win command fix gklab 2023-09-01 09:12:40 +08:00
  • 245a02efce
         adjust OpenBlas from cmake to make JackJollimore 2023-08-31 20:52:28 -03:00
  • 528134dd02
         remove convert-llama-7b-pth-to-gguf.py and convert-llama-hf-to-gguf.py (#2906) slaren 2023-09-01 01:32:09 +02:00
  • 6db44e09ce fix for #2940 staviq 2023-09-01 01:13:07 +02:00
  • 64ce7822bc make : use unaligned vector moves on MinGW Cebtenzzre 2023-08-31 18:54:22 -04:00
  • aeefac4ff7
         scripts: Use local gguf package when running from repo (#2927) Kerfuffle 2023-08-31 16:49:24 -06:00
  • 459d3e799d Just add local gguf to path unless environ forbids it KerfuffleV2 2023-08-31 15:26:09 -06:00
  • 2c17e6951a fix gpt-neox mmnga 2023-09-01 05:35:30 +09:00
  • 75939b41d5 cuda : vsubss4 for older versions of ROCm/clang Engininja2 2023-08-31 13:31:28 -06:00