Tags
Tags give the ability to mark specific points in history as being important
master-8c3ffc2
8c3ffc2f
·
ggml : update cblas_sgemm columns var to be more reasonable (#838)
·
Apr 13, 2023
master-585d91a
585d91a1
·
cmake : add explicit F16C option (x86) (#576)
·
Apr 13, 2023
master-95ea26f
95ea26f6
·
benchmark : add tool for timing q4_0 matrix multiplication (#653)
·
Apr 13, 2023
master-e7f6997
e7f6997f
·
Don't crash on ftype (formerly f16) == 4 (#917)
·
Apr 12, 2023
master-8b67998
8b679987
·
Fix whitespace, add .editorconfig, add GitHub workflow (#883)
·
Apr 11, 2023
master-3e6e70d
3e6e70d8
·
Add enum llama_ftype, sync ggml_type to model files (#709)
·
Apr 11, 2023
master-2663d2c
2663d2c6
·
Windows fixes (#890)
·
Apr 11, 2023
master-461ba9e
461ba9e6
·
ggml : fix WASM build
·
Apr 10, 2023
master-c3ac702
c3ac702e
·
ggml : add ggml_cont() + optimize ggml_cpy() for contiguous dst
·
Apr 10, 2023
master-d9a239c
d9a239c4
·
Simplify to include lower-case windows.h always, fix compile on mingw32 (#747)
·
Apr 10, 2023
master-684da25
684da259
·
ggml : fix quantize_row_q4_1() ARM_NEON (close #876)
·
Apr 10, 2023
master-180b693
180b693a
·
Print model version.
·
Apr 10, 2023
master-aaf3b23
aaf3b23d
·
fix for windows utf-8 input (#840)
·
Apr 08, 2023
master-f2d1c47
f2d1c472
·
cmake should link openblas properly with -lopenblas like how it's done in the makefile (#839)
·
Apr 08, 2023
master-62cfc54
62cfc54f
·
Add quantize-stats command for testing quantization (#728)
·
Apr 08, 2023
master-698f7b5
698f7b5d
·
make : add libllama.so target for llama-cpp-python (#797)
·
Apr 07, 2023
master-4953e90
4953e900
·
llama : always sort logits before nucleus sampling (#812)
·
Apr 07, 2023
master-cc9cee8
cc9cee8e
·
Do not crash when it has nothing to say. (#796)
·
Apr 06, 2023
master-eeaa7b0
eeaa7b04
·
ggml : multi-thread ggml_rope() (~3-4 times faster on M1) (#781)
·
Apr 05, 2023
master-986b6ce
986b6ce9
·
ggml, llama : avoid heavy V transpose + improvements (#775)
·
Apr 05, 2023
1
…
28
29
30
31
32
33
34
35
36
…
38