Commit graph

  • a3c1234bd5
    Merge branch 'ggerganov:master' into master m3ndax 2023-08-31 21:22:58 +02:00
  • 31fc2bdebe Life is suffering. KerfuffleV2 2023-08-31 12:37:56 -06:00
  • 4b54a7e7bc Look for local gguf module relative to running script file KerfuffleV2 2023-08-31 12:31:27 -06:00
  • 9df3427151 Prefix target with package name Mason M 2023-08-31 14:38:43 -03:00
  • e6425aef24 Add llama as import library Mason M 2023-08-31 14:31:23 -03:00
  • 0c3a265187 fixed incorrect buffer size values Concedo 2023-09-01 01:31:09 +08:00
  • 35ba699a7c Merge remote-tracking branch 'vxii/concedo' into concedo_experimental Concedo 2023-09-01 01:28:16 +08:00
  • f836df9043 Merge remote-tracking branch 'upstream/master' into fix-warnings Cebtenzzre 2023-08-31 13:26:11 -04:00
  • 0fe3c9cf96 stronger banning bias Concedo 2023-09-01 01:25:23 +08:00
  • 822127e6d1 Remove unnecessary conditional in addl token error handling KerfuffleV2 2023-08-31 11:24:46 -06:00
  • fe4a233d79 Merge branch 'master' into concedo_experimental Concedo 2023-09-01 00:47:06 +08:00
  • 25c0d79b0f convert.py: BPE fixes? KerfuffleV2 2023-08-31 10:46:56 -06:00
  • 7ea2c68831 Add RISC-V Vector Intrinsics Support Ahmad Tameem 2023-08-24 19:29:31 +05:00
  • f2985a070b
    Add support for 34B GGML models vxiiduu 2023-09-01 01:29:09 +10:00
  • ee1fd6024a docs: add java-llama.cpp to README.md Konstantin Herud 2023-08-31 17:01:45 +02:00
  • 4914f855c7
    add tensor checkpoints only when gradient checkpointing is enabled xaedes 2023-08-31 16:46:21 +02:00
  • e0da1684db
    remove finetune option to disable allocator xaedes 2023-08-31 16:45:47 +02:00
  • 9ea69b745e remove convert-llama-7b-pth-to-gguf.py and convert-llama-hf-to-gguf.py slaren 2023-08-30 16:43:52 +02:00
  • b0c6fa5656 Use proper build info Mason M 2023-08-31 09:30:24 -03:00
  • 93212d5371 Add logic to generate Config package Mason M 2023-08-31 09:18:28 -03:00
  • b860f65446 Allow quantize to only copy tensors, other improvements KerfuffleV2 2023-08-31 05:41:20 -06:00
  • e8422de39e
    @vxiiduu's fix for PrefetchVirtualMemory (#2930) b1132 DannyDaemonic 2023-08-31 04:21:45 -07:00
  • 18b15a6f49 @vxiiduu's fix for PrefetchVirtualMemory Co-authored-by: vxiiduu <73044267+vxiiduu@users.noreply.github.com> Danny Daemonic 2023-08-31 04:03:32 -07:00
  • 027930c0b7 remove unused GGML_NITER from workflows Alon Faraj 2023-08-31 13:59:46 +03:00
  • bc23fcd797 added support for RISCV CFLAGS & native compile + cross compile options moiz.hussain 2023-08-11 14:44:49 +05:00
  • bc02f7663f allow sse3 in failsafe Concedo 2023-08-31 18:07:17 +08:00
  • 489ffb2525 Only I could mess up something this simple. KerfuffleV2 2023-08-31 04:06:48 -06:00
  • 4320055b22 scripts: Use local gguf when running from repo KerfuffleV2 2023-08-31 04:01:17 -06:00
  • 12b611ace0 update make file -pb flag Alon Faraj 2023-08-31 12:53:56 +03:00
  • eb2b59c2f1 install lcov Alon Faraj 2023-08-31 12:28:32 +03:00
  • a589615fd3 - add CODECOV_TOKEN - add missing make lcov-report Alon Faraj 2023-08-31 12:19:25 +03:00
  • a62df95d35 add LLAMA_CODE_COVERAGE=1 again Alon Faraj 2023-08-31 12:16:23 +03:00
  • 5a4426c633 add env vars Alon Faraj 2023-08-31 12:14:27 +03:00
  • a5e2aa33f7 check why the job hang Alon Faraj 2023-08-31 12:11:15 +03:00
  • c4ced68344 use gcc-8 Alon Faraj 2023-08-31 12:02:03 +03:00
  • 89ff60e941 wun on ubuntu 20.04 Alon Faraj 2023-08-31 11:57:52 +03:00
  • 09e0f58692 update code coverage workflow Alon Faraj 2023-08-31 11:56:39 +03:00
  • 997a88c213 add code-coverage workflow Alon Faraj 2023-08-31 11:43:31 +03:00
  • 07b02af8bc fixed tab ordering , update lite for panel alignment Concedo 2023-08-31 16:33:00 +08:00
  • e2fd30b5d1 reverted the failsafe removal, since they dropped support for dll check Concedo 2023-08-31 15:39:32 +08:00
  • 92d0b751a7
    convert : fix python 3.8 support, modernize type annotations (#2916) Cebtenzzre 2023-08-31 01:02:23 -04:00
  • 35cd10c173 metal: yet another MUL mat-vec template lshzh-ww 2023-08-30 23:46:42 -04:00
  • d5d5c523c9 ggml : cleanup unused arm32 specific impl Jhen 2023-08-31 11:28:25 +08:00
  • aa4b7d29a2 metal: improvement for Q4_K driver lshzh-ww 2023-08-30 23:05:22 -04:00
  • ac671922f7 k-quants : fix build on armv7 Jhen 2023-08-31 09:22:55 +08:00
  • 32bbb554a7
    Made Usage example more easy to copy JackJollimore 2023-08-30 22:12:12 -03:00
  • 82700d8c38
    Update README.md JackJollimore 2023-08-30 21:55:25 -03:00
  • 2612b36061
    Update Usage Example JackJollimore 2023-08-30 21:53:26 -03:00
  • 397b3399cb fix more MinGW warnings Cebtenzzre 2023-08-30 20:05:05 -04:00
  • 41a58ce8da fix some MinGW warnings Cebtenzzre 2023-08-30 19:44:09 -04:00
  • a04a6990eb llama2c : use a destructor to prevent memory leaks Cebtenzzre 2023-08-30 18:32:13 -04:00
  • 634d7272b8 llama2c : fix basename on Windows Cebtenzzre 2023-08-30 18:22:52 -04:00
  • 78cd2d9c8a simplify win detection staviq 2023-08-31 00:18:35 +02:00
  • 4db4eb20d8 make LOG_COMPAT easier to override and extend staviq 2023-08-30 23:26:52 +02:00
  • 05df178874 gguf-py: bump version Cebtenzzre 2023-08-30 16:30:00 -04:00
  • 8a8c1cb0f2 convert : add necessary type: ignore comments Cebtenzzre 2023-08-30 17:04:55 -04:00
  • e66e0be432 add mypy.ini Cebtenzzre 2023-08-30 16:41:51 -04:00
  • c7cc756880
    metal : slight speed-up for add and mul kernels Georgi Gerganov 2023-08-30 23:22:40 +03:00
  • 42a0768c40 gguf.py : a tuple is already a tuple Cebtenzzre 2023-08-30 16:12:02 -04:00
  • ef87d46d68 convert : use PEP 585 generics and PEP 604 unions Cebtenzzre 2023-08-30 16:09:38 -04:00
  • f30937a5a2 convert : fix mypy errors in convert-llama-ggmlv3-to-gguf Cebtenzzre 2023-08-30 15:54:08 -04:00
  • 68738ff55a convert : fix required parameters in convert-llama-ggmlv3-to-gguf Cebtenzzre 2023-08-30 15:45:38 -04:00
  • a09b674a98 convert : sort imports Cebtenzzre 2023-08-30 15:37:16 -04:00
  • b18159b803 convert : fix python 3.8 support Cebtenzzre 2023-08-30 15:35:27 -04:00
  • 8afe228000
    CUDA: mul_mat_q=true llama_context_params default (#2912) b1130 Johannes Gäßler 2023-08-30 21:46:19 +02:00
  • 2cb9e7e312 Keep static libs and headers with install Mason M 2023-08-30 16:42:47 -03:00
  • 3a7e9eb817 Use stated vocab size akawrykow 2023-08-30 08:32:03 -07:00
  • a8c96f2cac llama2c : fix mismatch between new[] and delete Cebtenzzre 2023-08-30 14:46:22 -04:00
  • 8c2b881281 cuda : poc for norm quants (only -b 1 works) norm-quants Georgi Gerganov 2023-08-30 21:39:49 +03:00
  • 3985077331 llama2c : fix segfault if vocab is not found Cebtenzzre 2023-08-30 14:40:08 -04:00
  • ced231980e Remove warning which fails on windows. master-ced2319 Adam Treat 2023-08-30 14:33:31 -04:00
  • be1ddb142f CUDA: mul_mat_q=true llama_context_params default JohannesGaessler 2023-08-30 20:10:53 +02:00
  • f33eeadf76 formatting staviq 2023-08-30 20:10:02 +02:00
  • 353ed6e7c7 fix mingw-like builds staviq 2023-08-30 19:58:07 +02:00
  • ecf9af80ac Merge remote-tracking branch 'upstream/master' into fix-warnings Cebtenzzre 2023-08-30 13:40:58 -04:00
  • b6914ebd04 hotfix to revert the auto ctx scaling first, i didnt do it properly Concedo 2023-08-31 00:58:52 +08:00
  • df54d2f1d4
    ggml : use less ggml_mul tasks when src0 rows are few Georgi Gerganov 2023-08-30 19:37:26 +03:00
  • 1e85f6b711 Updated gitignore for new example. apaz-cli 2023-08-30 11:16:48 -05:00
  • 71d6975559
    [Docker] fix tools.sh argument passing. (#2884) Henri Vasserman 2023-08-30 19:14:53 +03:00
  • 16841ac0e8 Resolved merge conflicts. apaz-cli 2023-08-30 11:09:47 -05:00
  • 253eab8ae1
    ggml : poc for normalizing weights for better quantization (metal) Georgi Gerganov 2023-08-30 19:05:36 +03:00
  • 828a43d2b3 Added makefile, better error messages apaz-cli 2023-08-30 10:59:52 -05:00
  • 9129ba88f8 makefile: add coverage support (lcov, gcovr) Alon Faraj 2023-08-30 18:52:59 +03:00
  • f0e1978d38 update .gitignore Alon Faraj 2023-08-30 18:51:22 +03:00
  • b4e70822f6
    metal : add poc for normalized Q4_0 and Q4_1 norm-quants-rebase Georgi Gerganov 2023-08-30 18:32:43 +03:00
  • e3f10a992f
    Merge branch 'ggerganov:master' into master m3ndax 2023-08-30 17:35:30 +02:00
  • 4fd51c4616
    fix warnings xaedes 2023-08-30 17:12:23 +02:00
  • 5cd0309610 renamed incorrect identifier Concedo 2023-08-30 23:06:39 +08:00
  • 2e7c558d07 quick start command fix gklab 2023-08-30 23:02:21 +08:00
  • 4974f37054 Merge remote-tracking branch 'JohannesGaessler/skip-unused-2' into skip-unused-2 Olivier Chafik 2023-08-30 16:01:49 +01:00
  • d547e05f0e Merge remote-tracking branch 'origin/master' into skip-unused-2 Olivier Chafik 2023-08-30 16:00:44 +01:00
  • 0c57f9f0b3
    fix warnings xaedes 2023-08-30 16:55:49 +02:00
  • 9ffe54ed10
    Merge branch 'master' into norm-quants Georgi Gerganov 2023-08-30 16:26:59 +03:00
  • c436f18cc4 use mmq slaren 2023-08-30 16:42:01 +02:00
  • 0ee394ae1b falcon disable offload only for clblast Concedo 2023-08-30 22:35:24 +08:00
  • 4e986ac4bc
    update README.md xaedes 2023-08-30 16:29:09 +02:00
  • b26bd4c34c
    add option to save train-text-from-scratch output every N iterations xaedes 2023-08-30 16:26:05 +02:00
  • bb0cadf678 New python based entry point for containers Jiri Podivin 2023-05-28 16:38:03 +02:00
  • 29757de61f cmake disable buggy logs Concedo 2023-08-30 22:15:33 +08:00
  • 4cdaa3c9cb Nomic vulkan backend licensed under the Software for Open Models License (SOM), version 1.0. niansa 2023-06-22 12:58:07 +02:00