vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
440 commits 380 branches 3056 tags 365 MiB
Commit graph

440 commits

This branch
This branch
All branches
Author SHA1 Message Date
ajasibley
3abc987f2a
Merge pull request #9 from plurigrid/babashka 
Cleaned up shell.nix by removing bash commands and replacing them with just recipes
 2023-05-26 03:11:41 -07:00
ajasibley
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. 2023-05-25 12:23:07 -07:00
ajasibley
207389d91a Added support in update_source.sh for device with no zshrc made. 2023-05-25 03:15:36 -07:00
ajasibley
d54795176c Fixed syntac error in update_source.sh. 2023-05-25 02:46:49 -07:00
ajasibley
943ad39016 Fixed typo cuasing path shell source to not be reloaded in justfile. 2023-05-25 02:41:03 -07:00
ajasibley
fa8bc623c8 Fixed typo in install-cosmo just recipe. 2023-05-25 02:32:20 -07:00
ajasibley
6cdc1f6aa9 Updated installation check in install-rust just recipe. 2023-05-25 02:27:08 -07:00
ajasibley
90c750a6ad Added automatic path setting to install-cosmo just recipe. 2023-05-25 02:24:43 -07:00
ajasibley
61cd6685f4 Fixed syntax error in update_path.sh 2023-05-25 02:13:01 -07:00
ajasibley
28491b7515 fixed syntax error in install_nix.sh. 2023-05-25 02:04:14 -07:00
ajasibley
1c1245535a fixed syntax error in nix just recipe. 2023-05-25 01:58:07 -07:00
ajasibley
5bd0c41792 Added automatic Nix install if not present on system. 2023-05-25 01:47:57 -07:00
ajasibley
0f4c13fc43 Updated justfile, added script to set shell path, and fixed Cargo.toml version issue. 2023-05-25 01:19:57 -07:00
ajasibley
875b385d79
Update README.md 2023-05-24 17:37:37 -07:00
ajasibley
9580a547d9 Updated README. 2023-05-24 16:56:17 -07:00
ajasibley
5d66c80d99 Cleaner up shell.nix by removing bash commands and replacing them with just recipes. 2023-05-24 15:07:33 -07:00
ajasibley
c0e8da7912
Update README.md 2023-05-24 04:41:50 +00:00
ajasibley
e8ef92a738
Update README.md 2023-05-24 04:30:05 +00:00
ajasibley
988dd73d82
Merge pull request #7 from plurigrid/cosmonic 
Removed Vespa project
 2023-05-24 04:28:26 +00:00
Aja Sibley
2df76d9064 Removed Vespa project. 2023-05-24 04:26:59 +00:00
barton ⊛
74ac17d3ad
Merge pull request #5 from plurigrid/cosmonic 
Cosmonic wasmCloud via nix shell, improving dev experience: tested on aarch64
 2023-05-24 04:11:22 +00:00
ajasibley
025f974d68 Updated readme 2023-05-23 20:24:13 -07:00
ajasibley
5f727081bc Fixed env error 2023-05-23 19:15:11 -07:00
ajasibley
3d6d096ab8 Update shell.nix 
Added Cosmonic to path
 2023-05-23 23:10:25 +00:00
ajasibley
b45efa3239 Updated nix.shell 
Added install and setup for Cosmonic and its dependencies for Apple Silicon Macs and and Nvidia Jetsons.
 2023-05-23 22:20:17 +00:00
ajasibley
9a565fc9a1
Update README.md 2023-05-18 12:39:44 +00:00
ajasibley
88b988e74b
Update README.md 2023-05-18 12:38:18 +00:00
ajasibley
2dd8d16948
Update README.md 2023-05-18 12:38:06 +00:00
ajasibley
3c0ab1f8b8
Update README.md 2023-05-18 12:37:35 +00:00
ajasibley
954dccfd22
Update README.md 2023-05-18 12:37:05 +00:00
Aja Sibley
4905d355b5 Updated shell nix to install cosmo if not already installed. Added README and cleaned up repo. 2023-05-18 12:36:09 +00:00
Barton Rhodes
119bddaa32 authenticity is all you need 2023-05-18 03:22:13 +00:00
Aja Sibley
26e842a342 Added error log for cosmo. 2023-05-18 01:07:10 +00:00
Aja Sibley
985ab154ec Added openssl to nix for cosmonic. 2023-05-18 00:07:39 +00:00
Aja Sibley
c3ad27c5df Updated comments. 2023-05-17 21:19:50 +00:00
Aja Sibley
933bb643a7 Fixed the vespa overlay. 2023-05-17 21:18:23 +00:00
Aja Sibley
19e8d6e288 Added vesp overlay. 2023-05-17 20:52:11 +00:00
Barton Rhodes
7c36e03dfb too many hooved animals! 2023-04-21 05:07:57 +00:00
Barton Rhodes
6ffe4680ca fiat nexus ▦ 2023-04-20 22:38:56 +00:00
barton ⊛
041627284d
Merge branch 'ggerganov:master' into main 2023-04-20 22:21:31 +00:00
Stephan Walter
2510c1831f
Add ggml-model-*.bin checksums for 7B, 13B, 30B, 65B (#1088) 
* Add ggml-model-*.bin checksums for 7B, 13B, 30B
* Add ggml-model-*.bin checksums for 65B

---------

Co-authored-by: Pavol Rusnak <pavol@rusnak.io>
 2023-04-20 23:56:44 +02:00
Georgi Gerganov
12b5900dbc
ggml : sync ggml (add GPT-NeoX RoPE implementation) 2023-04-20 23:32:59 +03:00
Georgi Gerganov
9ff334f3c9
ggml : fix bug in ggml_compute_forward_dup_f32() 2023-04-20 21:58:38 +03:00
slaren
2005469ea1
Add Q4_3 support to cuBLAS (#1086) 2023-04-20 20:49:53 +02:00
Georgi Gerganov
8a1756abdf
ggml : do not break cuBLAS build (Q4_3 is not yet implemented) 2023-04-20 21:43:50 +03:00
Georgi Gerganov
66aab46079
ggml : fix Q4_3 quantization 
Broke it during conflict resolution in last PR
 2023-04-20 20:44:05 +03:00
Kawrakow
38de86a711
llama : multi-threaded quantization (#1075) 
* Multi-threading quantization.

Not much gain for simple quantizations, bit it will be important
for quantizations that require more CPU cycles.

* Multi-threading for quantize-stats

It now does the job in ~14 seconds on my Mac for
Q4_0, Q4_1 and Q4_2. Single-threaded it was taking
more than 2 minutes after adding the more elaborate
version of Q4_2.

* Reviewer comments

* Avoiding compiler confusion

After changing chunk_size to const int as suggested by
@ggerganov, clang and GCC starting to warn me that I don't
need to capture it in the lambda. So, I removed it from the
capture list. But that makes the MSVC build fail. So,
making it a constexpr to make every compiler happy.

* Still fighting with lambda captures in MSVC

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
 2023-04-20 20:42:27 +03:00
Georgi Gerganov
e0305ead3a
ggml : add Q4_3 quantization (#1082) 2023-04-20 20:35:53 +03:00
Ivan Komarov
6a9661ea5a
ci : remove the LLAMA_ACCELERATE matrix dimension from Ubuntu builds in the CI (#1074) 
[Accelerate](https://developer.apple.com/documentation/accelerate) is an Apple framework which can only be used on macOS, and the CMake build [ignores](https://github.com/ggerganov/llama.cpp/blob/master/CMakeLists.txt#L102) the `LLAMA_ACCELERATE` variable when run on non-Apple platforms. This implies setting `LLAMA_ACCELERATE` is a no-op on Ubuntu and can be removed.

This will reduce visual noise in CI check results (in addition to reducing the number of checks we have to run for every PR). Right now every sanitized build is duplicated twice for no good reason (e.g., we have `CI / ubuntu-latest-cmake-sanitizer (ADDRESS, Debug, ON)` and `CI / ubuntu-latest-cmake-sanitizer (ADDRESS, Debug, OFF)`).
 2023-04-20 18:15:18 +03:00
源文雨
5addcb120c
fix: LLAMA_CUBLAS=1 undefined reference 'shm_open' (#1080) 2023-04-20 15:28:43 +02:00