Iwan Kawrakow 70815ec5b2 iq2k: quantization improvement ...

I was not using the ciorrect scale sign to compute
mse when checking the solution with the sign flipped.
iq4_kss is now almost on par with the 4-bit Trellis.

2024-11-24 11:29:37 +02:00

..

cmake

Merge mainline llama.cpp (#3)

2024-07-27 07:55:01 +02:00

include

Adding iq4_kt - not competitive at this point

2024-11-21 08:16:41 +02:00

src

iq2k: quantization improvement

2024-11-24 11:29:37 +02:00

.gitignore

Merge mainline llama.cpp (#3)

2024-07-27 07:55:01 +02:00

CMakeLists.txt

Move to c++17 projectwide (#80)

2024-10-04 14:43:26 +03:00