char-fork/llama.cpp

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

Djip007ggml : refactor online repacking (#10446)19d8762ab

char/aliceai-t5-moe
brancheschar/aliceai-t5-moeall refs →
397 B10 linesraw
1#pragma once
2#include "common.h"
3
4size_t ggml_backend_amx_desired_wsize(const struct ggml_tensor * dst);
5
6size_t ggml_backend_amx_get_alloc_size(const struct ggml_tensor * tensor);
7
8void ggml_backend_amx_convert_weight(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size);
9
10void ggml_backend_amx_mul_mat(const struct ggml_compute_params * params, struct ggml_tensor * dst);