 rankaiyx
|
c238b5873a
benchmark-matmul: Print the average of the test results (#1490)
|
пре 2 година |
 Tom Jobbins
|
2b2646931b
convert.py: Support models which are stored in a single pytorch_model.bin (#1469)
|
пре 2 година |
 Ilya Kurdyukov
|
42627421ec
~7% faster Q5_1 AVX2 code (#1477)
|
пре 2 година |
 András Salamon
|
9560655409
define default model path once, sync path with readme (#1366)
|
пре 2 година |
 sandyiscool
|
2a5ee023ad
Add alternate include path for openblas (#1476)
|
пре 2 година |
 zrm
|
63d20469b8
fix get_num_physical_cores() (#1436)
|
пре 2 година |
 slaren
|
b5c9295eef
benchmark-matmul: fix clang-tidy issues, report results in GFLOPS (#1458)
|
пре 2 година |
 Johannes Gäßler
|
eb363627fd
cuda : deduplicated dequantization code (#1453)
|
пре 2 година |
 xaedes
|
79b2d5b69d
ggml : alternative fix for race condition bug in non-inplace ggml_compute_forward_diag_mask_f32 (#1454)
|
пре 2 година |
 Georgi Gerganov
|
13c351ad72
ggml : various fixes (#1450)
|
пре 2 година |
 katsu560
|
60f8c361ca
ggml : add AVX support based on AVX2 code (#1430)
|
пре 2 година |
|
Georgi Gerganov
|
601a033475
ggml : add GGML_QNT_VERSION to track quantization format changes
|
пре 2 година |
|
Georgi Gerganov
|
08737ef720
cuda : fix convert function (#1412)
|
пре 2 година |
|
Georgi Gerganov
|
bda4d7c215
make : fix PERF build with cuBLAS
|
пре 2 година |
|
Georgi Gerganov
|
5a5aeb1e91
llama : fix unused warning
|
пре 2 година |
|
Georgi Gerganov
|
66841fdb0e
ggml : multi-thread mul and diag_mask ops (#1428)
|
пре 2 година |
|
Johannes Gäßler
|
905d87b70a
ggml : GPU-accelerated token generation (#1412)
|
пре 2 година |
|
xaedes
|
f954edda93
ggml : implement backward pass for llama + small training-llama-from-scratch example (#1360)
|
пре 2 година |
|
Georgi Gerganov
|
f048af0230
ggml : sync alibi fix from ggml repo
|
пре 2 година |
 3ooabkhxtn
|
ac0cd259d5
Adding SSE instructions to ggml_vec_dot_q4_0_q8_0 (#1413)
|
пре 2 година |
|
Georgi Gerganov
|
0cd22e190a
llama : fix various warnings
|
пре 2 година |
 Rinne
|
6456a4eb9f
embedding : remove unused code (#1426)
|
пре 2 година |
|
Georgi Gerganov
|
cdd5350892
readme : update Q4_0 perplexities
|
пре 2 година |
|
Georgi Gerganov
|
738ace394a
llama : free ggml context in set / copy state data (close #1425)
|
пре 2 година |
 Henri Vasserman
|
699b1ad7fe
opencl : fix kernels for the new formats (#1422)
|
пре 2 година |
|
Georgi Gerganov
|
fb62f92433
llama : fix --mtest option (close #1414)
|
пре 2 година |
|
Johannes Gäßler
|
773ee249fb
CLI args use - instead of _, backwards compatible (#1416)
|
пре 2 година |
|
slaren
|
553fd4d4b5
Add clang-tidy reviews to CI (#1407)
|
пре 2 година |
 Rinne
|
089b1c93ba
readme : add C#/.NET bindings repo (#1409)
|
пре 2 година |
|
Georgi Gerganov
|
b9fd7eee57
ggml : remove bit shuffling (#1405)
|
пре 2 година |
