Francis Couture-Harpin e9719576c4 ggml : also faster TQ1_0 ... Same optimization as for TQ2_0 by offsetting the sum instead of the weights. This makes TQ1_0 almost as fast as Q8_0 on AVX2.		2024-07-31 00:08:48 -04:00
..
cmake	llama : reorganize source code + improve CMake (#8006)	2024-06-26 18:33:02 +03:00
include	ggml : add TQ1_0 and TQ2_0 ternary quantization types	2024-07-30 18:33:15 -04:00
src	ggml : also faster TQ1_0	2024-07-31 00:08:48 -04:00
.gitignore	vulkan : cmake integration (#8119)	2024-07-13 18:12:39 +02:00
CMakeLists.txt	feat: Support Moore Threads GPU (#8383)	2024-07-28 01:41:25 +02:00