Commit graph

  • 05cb629c8e
    replace inefficient repeat backward pass with dedicated repeat_back operation xaedes 2023-05-28 18:00:17 +02:00
  • c47df09842
    simplify backward pass for SQRT xaedes 2023-05-28 17:32:01 +02:00
  • 15ddc4903b Merge remote-tracking branch 'slyecho/server_refactor' digiwombat 2023-05-28 11:09:32 -04:00
  • 36758b1009 Setting the ftype argument of the script as optional Jiri Podivin 2023-05-28 16:39:51 +02:00
  • 7186d655a1
    seed and gen params Henri Vasserman 2023-05-28 17:03:01 +03:00
  • 7740301db9 Set unspecified generation settings back to default. (Notes below) digiwombat 2023-05-28 09:18:47 -04:00
  • dda915cac4 Added capturing the stopping word and sending it along with the final JSON. digiwombat 2023-05-28 08:43:38 -04:00
  • 2e5c5ee224 Changed JSON names to match the parameter name rather than the variable name. digiwombat 2023-05-28 08:12:48 -04:00
  • 23928f2887 Added generation_settings to final json object. digiwombat 2023-05-28 08:04:05 -04:00
  • 5eacb84223 Display a warning if -ngl is supplied without support. KerfuffleV2 2023-05-28 05:48:36 -06:00
  • e8efd75492 Initial timeout code and expanded json return on completion. digiwombat 2023-05-28 07:44:31 -04:00
  • 28f1196f65 adjust default rep pen range Concedo 2023-05-28 19:36:21 +08:00
  • 177868e68a Changed to params/args digiwombat 2023-05-28 06:29:11 -04:00
  • a70095e961 Fix derp in ngl ifdef KerfuffleV2 2023-05-28 04:17:44 -06:00
  • f40f6e8252 Documentation and arg help/handling updates KerfuffleV2 2023-05-28 03:58:03 -06:00
  • 764a21ce0f Only show -ngl option when relevant + add warning for --memory-f32 option KerfuffleV2 2023-05-28 03:36:52 -06:00
  • 549291fe61
    keep processed from the beginning Henri Vasserman 2023-05-28 12:08:37 +03:00
  • df0e0d094c
    Forgot to remove some testing code. Randall Fitzgerald 2023-05-23 06:22:30 -07:00
  • f93fe36c5b
    Add all generation parameters to server.cpp and allow resetting context Randall Fitzgerald 2023-05-23 06:16:54 -07:00
  • 51e09944ce
    server rewrite Henri Vasserman 2023-05-28 02:42:18 +03:00
  • c01c7d2caf Adding git in container package dependencies Jiri Podivin 2023-05-28 09:33:33 +02:00
  • 7d159bacd7 updated kobold lite Concedo 2023-05-28 11:23:20 +08:00
  • 0d308e2ef2 remove excessive codes and prints liang 2023-05-28 08:45:51 +08:00
  • 1f40a789e6
    Didn't see the already defined top_k var. Randall Fitzgerald 2023-05-27 17:10:09 -07:00
  • e84b802161
    Change top_k type. Randall Fitzgerald 2023-05-27 17:07:45 -07:00
  • fdce8951ac
    Merge branch 'ggerganov:master' into master Randall Fitzgerald 2023-05-27 19:57:37 -04:00
  • d20f36b93c
    Removed unnecessary last_prompt_token set Randall Fitzgerald 2023-05-27 16:46:05 -07:00
  • 36c86d794d
    Automate Context resetting and minor fixes Randall Fitzgerald 2023-05-27 16:43:08 -07:00
  • dbf50b4afc
    Merge 95dc4d7270 into a6704643b6 Henri Vasserman 2023-05-27 13:51:52 -07:00
  • a6704643b6
    ggml : add support for the RISCV architecture (#1616) master-a670464 apcameron 2023-05-27 21:03:25 +01:00
  • 6dea8d9f2e LLAMA_DEBUG adds debug symbols JohannesGaessler 2023-05-27 21:28:08 +02:00
  • 66ed19d01f
    Corrected dashes in the help lines. Randall Fitzgerald 2023-05-27 11:51:21 -07:00
  • dcc426e2de Merge branch 'master' into concedo_experimental Concedo 2023-05-28 01:08:39 +08:00
  • 48cb16a51a
    Merge branch 'ggerganov:master' into master Randall Fitzgerald 2023-05-27 13:08:03 -04:00
  • 0df7d63e5b
    Include server in releases + other build system cleanups (#1610) master-0df7d63 Kerfuffle 2023-05-27 11:04:14 -06:00
  • 5d9f5b28a6 rwkv integration completed Concedo 2023-05-28 00:48:56 +08:00
  • 9fdaa1d250
    Add more defs Henri Vasserman 2023-05-27 19:17:53 +03:00
  • 81996ea82f Minor style fix in cast KerfuffleV2 2023-05-27 06:10:51 -06:00
  • 2c5f38ba92 Apply style/style suggestions for server example KerfuffleV2 2023-05-27 06:00:26 -06:00
  • b70917d2b4 Include server in releases, other build system cleanups KerfuffleV2 2023-05-27 05:08:58 -06:00
  • 97c9b77c4f
    Add documentation about CLBlast (#1604) Henri Vasserman 2023-05-27 18:47:55 +03:00
  • a4648c1e7c
    Merge 'origin/master' into hipblas Henri Vasserman 2023-05-27 18:22:39 +03:00
  • 55e0fbf024 wip integrating new rwkv Concedo 2023-05-27 22:45:28 +08:00
  • 98edbdba63 ADD Support for the RISCV Architecture Andrew Cameron 2023-05-27 15:34:40 +01:00
  • a4e5f65659 Ensure ext_buffer is null terminated Howard Su 2023-05-27 21:04:42 +08:00
  • 6c5c79a2a3 Use strstr to check if fp16 supported Howard Su 2023-05-27 20:13:49 +08:00
  • 8c1a3746ba No need to allocate cl_mem on heap Howard Su 2023-05-27 20:17:41 +08:00
  • 605f3fe290
    Add --alias option to gpt_params to set use friendly model name Vladimir Zorin 2023-05-27 17:25:37 +03:00
  • 0ecb1bbbeb
    [CI] Fix openblas (#1613) master-0ecb1bb Henri Vasserman 2023-05-27 17:24:06 +03:00
  • e84e30db2f
    fiiixx Henri Vasserman 2023-05-27 16:56:46 +03:00
  • 4be5c264a7
    fix license copy Henri Vasserman 2023-05-27 16:40:34 +03:00
  • 42d8271ee3
    syntax error fix Henri Vasserman 2023-05-27 16:24:03 +03:00
  • 699bd99d4e
    Fix openblas build Henri Vasserman 2023-05-27 16:22:44 +03:00
  • 93618031c7
    ggml : add ggml_tensor_overhead() master-9361803 Georgi Gerganov 2023-05-27 16:19:56 +03:00
  • 83c54e6da5
    [CI] CLBlast: Fix directory name (#1606) Henri Vasserman 2023-05-27 15:18:25 +03:00
  • 6d47258e41 n_session_consumed should just be size_t, but the cache code casts to (int) Danny Daemonic 2023-05-27 03:36:44 -07:00
  • 573b690e16 Work around for recalculating logits in cached prompts Danny Daemonic 2023-05-27 03:20:03 -07:00
  • fe63bfdb0f Revert "allow 2048 blasbatchsize" Concedo 2023-05-27 18:13:27 +08:00
  • 97c5cca4e5 OpenCL: Don't load gpu layers into RAM, add mul_f32 kernel 0cc4m 2023-05-27 12:00:56 +02:00
  • 94dc5c2324 allow 2048 blasbatchsize Concedo 2023-05-27 17:47:18 +08:00
  • 92a0d77712 Merge branch 'master' into concedo_experimental Concedo 2023-05-27 17:44:14 +08:00
  • abfdfb702e added top_a sampler Concedo 2023-05-27 17:32:37 +08:00
  • bdbda1b17a
    ggml : sync ggml core (minor additions, e.g. ggml_get_tensor_by_name()) Georgi Gerganov 2023-05-27 12:22:05 +03:00
  • ebc5d0651a Use events instead of clFinish, where possible 0cc4m 2023-05-27 10:03:35 +02:00
  • 01a0f206df added support for starcoder, which is basically gpt2 Concedo 2023-05-27 13:35:40 +08:00
  • 6d7749c98f no difference Concedo 2023-05-27 12:42:19 +08:00
  • bd4fe936f5 cleanup sampling code Concedo 2023-05-27 11:58:39 +08:00
  • 20054a38c1
    Fix directory name fix_clblast Henri Vasserman 2023-05-27 02:00:08 +03:00
  • 7be1bbead5 Updated vespa-cli version. Aja Sibley 2023-05-26 22:54:00 +00:00
  • 20dd97c9a9
    fix trailing ws Henri Vasserman 2023-05-27 00:37:31 +03:00
  • 77f8d1d3b8
    Add documentation about CLBlast Henri Vasserman 2023-05-27 00:30:31 +03:00
  • ff348be22e Consistently catch and throw only exceptions deriving from std::exception mgroeber9110 2023-05-26 12:29:58 +02:00
  • 3abc987f2a
    Merge pull request #9 from plurigrid/babashka ajasibley 2023-05-26 03:11:41 -07:00
  • f519e81995 convert.py: add mapping for safetensors bf16 Aarni Koskela 2023-05-26 13:02:05 +03:00
  • 3c8f404243 integrated token probability viewer in debugmode Concedo 2023-05-26 16:40:26 +08:00
  • 991002b28e add some info in loading Liu Ming 2023-05-26 16:18:49 +08:00
  • 66874d4fbc
    Some improvements to loading the session with --prompt-cache (#1550) master-66874d4 Kerfuffle 2023-05-25 20:18:01 -06:00
  • 4c8b3fb107
    add configurable vars Henri Vasserman 2023-05-26 01:08:53 +03:00
  • 30d921af3e
    and makefile Henri Vasserman 2023-05-26 01:03:56 +03:00
  • a593a4f6c2
    Add missing parameters Henri Vasserman 2023-05-26 00:55:28 +03:00
  • 174bf6a86d
    Merge 'origin/master' into hipblas Henri Vasserman 2023-05-26 00:44:23 +03:00
  • 1fcdcc28b1
    cuda : performance optimizations (#1530) master-1fcdcc2 Johannes Gäßler 2023-05-25 23:07:29 +02:00
  • c2b55cc917
    Added LoRA Loading Randall Fitzgerald 2023-05-25 12:53:05 -07:00
  • 19d318946b Updated install_openssl.sh to properly set homebrew shell profile. Added message to just file promting user to restart shell following sucefull installation. Updated README to include steps for loging in to cosmonic and launching wasmCloust host. ajasibley 2023-05-25 12:23:07 -07:00
  • e09c67d141 Merge branch 'master' of github.com:liangmanlai/llama.cpp liang 2023-05-25 23:08:30 +08:00
  • 0fc6170173 In the function , add the cuda error bypass. liang 2023-05-25 23:03:14 +08:00
  • 207389d91a Added support in update_source.sh for device with no zshrc made. ajasibley 2023-05-25 03:15:36 -07:00
  • d54795176c Fixed syntac error in update_source.sh. ajasibley 2023-05-25 02:46:49 -07:00
  • 943ad39016 Fixed typo cuasing path shell source to not be reloaded in justfile. ajasibley 2023-05-25 02:41:03 -07:00
  • fa8bc623c8 Fixed typo in install-cosmo just recipe. ajasibley 2023-05-25 02:32:20 -07:00
  • 6cdc1f6aa9 Updated installation check in install-rust just recipe. ajasibley 2023-05-25 02:27:08 -07:00
  • 90c750a6ad Added automatic path setting to install-cosmo just recipe. ajasibley 2023-05-25 02:24:43 -07:00
  • aeacc57d3b
    Fixes for model conversion Henri Vasserman 2023-05-25 12:15:33 +03:00
  • 4905366fe8 Merge branch 'master' of github.com:baijiafan/llama.cpp Liu Ming 2023-05-25 17:13:56 +08:00
  • 18b778b967 update Liu Ming 2023-05-25 17:13:50 +08:00
  • 61cd6685f4 Fixed syntax error in update_path.sh ajasibley 2023-05-25 02:13:01 -07:00
  • 1fdd8ac615 update grpc impl 刘铭 2023-05-25 17:12:26 +08:00
  • 28491b7515 fixed syntax error in install_nix.sh. ajasibley 2023-05-25 02:04:14 -07:00
  • 1c1245535a fixed syntax error in nix just recipe. ajasibley 2023-05-25 01:58:07 -07:00
  • 5bd0c41792 Added automatic Nix install if not present on system. ajasibley 2023-05-25 01:47:57 -07:00