char-fork/llama.cpp

git clone https://git.t4t.associates/char-fork/llama.cpp

Akarshan Biswassycl : add flash-attn support for head size 512 (#21654)b54cb2e3d

char/aliceai-t5-moe
brancheschar/aliceai-t5-moeall refs →
345 B8 linesraw
1// This file has been autogenerated by generate_cu_files.py, do not edit manually.
2
3#include "../fattn-vec.hpp"
4
5DECL_FATTN_VEC_CASE( 64, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0);
6DECL_FATTN_VEC_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0);
7DECL_FATTN_VEC_CASE(256, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0);
8DECL_FATTN_VEC_CASE(512, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0);