 Georgi Gerganov
|
3a007648f2
metal : add option to disable debug logs (close #2764)
|
%!s(int64=2) %!d(string=hai) anos |
 Henri Vasserman
|
6bbc598a63
ROCm Port (#1087)
|
%!s(int64=2) %!d(string=hai) anos |
|
Georgi Gerganov
|
6381d4e110
gguf : new file format with flexible meta data (beta) (#2398)
|
%!s(int64=2) %!d(string=hai) anos |
 Kolen Cheung
|
0919a0f73d
cmake : install ggml-meta.metal if LLAMA_METAL (#2449)
|
%!s(int64=2) %!d(string=hai) anos |
 Shouzheng Liu
|
bf83bff674
metal : matrix-matrix multiplication kernel (#2615)
|
%!s(int64=2) %!d(string=hai) anos |
 Johannes Gäßler
|
f64d44a9b9
CUDA: Fixed OpenLLaMA 3b mmq, reduced compile time (#2590)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
4f6b60c776
CUDA: Fix models with output size != 32000 (#2480)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
b772bba42e
CUDA: fixed cmake F16 option (#2471)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
0728c5a8b9
CUDA: mmq CLI option, fixed mmq build issues (#2453)
|
%!s(int64=2) %!d(string=hai) anos |
 slaren
|
a113689571
ggml : add graph tensor allocator (#2411)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
11f3ca06b8
CUDA: Quantized matrix matrix multiplication (#2160)
|
%!s(int64=2) %!d(string=hai) anos |
 Cebtenzzre
|
6df1f5940f
make : build with -Wmissing-prototypes (#2394)
|
%!s(int64=2) %!d(string=hai) anos |
 wzy
|
78a3d13424
flake : remove intel mkl from flake.nix due to missing files (#2277)
|
%!s(int64=2) %!d(string=hai) anos |
|
wzy
|
45a1b07e9b
flake : update flake.nix (#2270)
|
%!s(int64=2) %!d(string=hai) anos |
|
wzy
|
b1f4290953
cmake : install targets (#2256)
|
%!s(int64=2) %!d(string=hai) anos |
 Howard Su
|
4e7464ef88
FP16 is supported in CM=6.0 (#2177)
|
%!s(int64=2) %!d(string=hai) anos |
 Evan Miller
|
5656d10599
mpi : add support for distributed inference via MPI (#2099)
|
%!s(int64=2) %!d(string=hai) anos |
 clyang
|
3bbc1a11f0
ggml : fix buidling with Intel MKL but ask for "cblas.h" issue (#2104) (#2115)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
924dd22fd3
Quantized dot products for CUDA mul mat vec (#2067)
|
%!s(int64=2) %!d(string=hai) anos |
 Tobias Lütke
|
7ee76e45af
Simple webchat for server (#1998)
|
%!s(int64=2) %!d(string=hai) anos |
 Daniel Drake
|
b213227067
cmake : don't force -mcpu=native on aarch64 (#2063)
|
%!s(int64=2) %!d(string=hai) anos |
 Kawrakow
|
6769e944c7
k-quants : support for super-block size of 64 (#2001)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
bbca06e269
cmake: revert CUDA arch default to 52, 61 if f16 (#1959)
|
%!s(int64=2) %!d(string=hai) anos |
|
Georgi Gerganov
|
23fc5c219a
cmake : fix trailing whitespaces
|
%!s(int64=2) %!d(string=hai) anos |
|
Howard Su
|
1e3abfcef0
cmake : fix build shared ggml when CUDA is enabled (#1929)
|
%!s(int64=2) %!d(string=hai) anos |
|
Johannes Gäßler
|
16b9cd1939
Convert vector to f16 for dequantize mul mat vec (#1913)
|
%!s(int64=2) %!d(string=hai) anos |
|
Howard Su
|
57cd69460f
cmake : add CUDA_ARCHITECTURES to new target ggml_static (#1917)
|
%!s(int64=2) %!d(string=hai) anos |
 Kerfuffle
|
b4c6f46f17
Allow cmake to build ggml as a library (#1896)
|
%!s(int64=2) %!d(string=hai) anos |
 Zenix
|
13fe9d2d84
cmake : add auto detection of BLAS_INCLUDE_DIRS (#1886)
|
%!s(int64=2) %!d(string=hai) anos |
|
Kawrakow
|
3d01122610
CUDA : faster k-quant dot kernels (#1862)
|
%!s(int64=2) %!d(string=hai) anos |
