char-fork/llama.cpp

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

Neo Zhang Jianyu[SYCL] refactor soft_max, add soft_max_back (#16472)b26021375

char/aliceai-t5-moe
brancheschar/aliceai-t5-moeall refs →
622 B24 linesraw
1//
2// MIT license
3// Copyright (C) 2024 Intel Corporation
4// SPDX-License-Identifier: MIT
5//
6
7//
8// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
9// See https://llvm.org/LICENSE.txt for license information.
10// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
11//
12
13#ifndef GGML_SYCL_SOFTMAX_HPP
14#define GGML_SYCL_SOFTMAX_HPP
15
16#include "common.hpp"
17
18#define SYCL_SOFT_MAX_BLOCK_SIZE 1024
19
20void ggml_sycl_op_soft_max(ggml_backend_sycl_context &ctx, ggml_tensor *dst);
21
22void ggml_sycl_op_soft_max_back(ggml_backend_sycl_context & ctx, ggml_tensor * dst);
23
24#endif // GGML_SYCL_SOFTMAX_HPP