Tags
Tags give the ability to mark specific points in history as being important
master-c238b58
c238b587
·
benchmark-matmul: Print the average of the test results (#1490)
·
May 17, 2023
master-4262742
42627421
·
~7% faster Q5_1 AVX2 code (#1477)
·
May 16, 2023
master-9560655
95606554
·
define default model path once, sync path with readme (#1366)
·
May 16, 2023
master-2a5ee02
2a5ee023
·
Add alternate include path for openblas (#1476)
·
May 16, 2023
master-63d2046
63d20469
·
fix get_num_physical_cores() (#1436)
·
May 15, 2023
master-b5c9295
b5c9295e
·
benchmark-matmul: fix clang-tidy issues, report results in GFLOPS (#1458)
·
May 14, 2023
master-79b2d5b
79b2d5b6
·
ggml : alternative fix for race condition bug in non-inplace...
·
May 14, 2023
master-13c351a
13c351ad
·
ggml : various fixes (#1450)
·
May 14, 2023
master-601a033
601a0334
·
ggml : add GGML_QNT_VERSION to track quantization format changes
·
May 14, 2023
master-bda4d7c
bda4d7c2
·
make : fix PERF build with cuBLAS
·
May 13, 2023
master-5a5aeb1
5a5aeb1e
·
llama : fix unused warning
·
May 13, 2023
master-66841fd
66841fdb
·
ggml : multi-thread mul and diag_mask ops (#1428)
·
May 13, 2023
master-905d87b
905d87b7
·
ggml : GPU-accelerated token generation (#1412)
·
May 13, 2023
master-f954edd
f954edda
·
ggml : implement backward pass for llama + small training-llama-from-scratch example (#1360)
·
May 13, 2023
master-f048af0
f048af02
·
ggml : sync alibi fix from ggml repo
·
May 13, 2023
master-ac0cd25
ac0cd259
·
Adding SSE instructions to ggml_vec_dot_q4_0_q8_0 (#1413)
·
May 13, 2023
master-0cd22e1
0cd22e19
·
llama : fix various warnings
·
May 13, 2023
master-6456a4e
6456a4eb
·
embedding : remove unused code (#1426)
·
May 13, 2023
master-738ace3
738ace39
·
llama : free ggml context in set / copy state data (close #1425)
·
May 13, 2023
master-699b1ad
699b1ad7
·
opencl : fix kernels for the new formats (#1422)
·
May 13, 2023
1
…
19
20
21
22
23
24
25
26
27
…
38