mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
9477661f4c
support ring attn for bert_attention plugin and dit model Signed-off-by: ChunhuanLin <lch_xdu@163.com>
35 lines
1.0 KiB
C++
35 lines
1.0 KiB
C++
/*
|
|
* Copyright (c) 2020-2024, NVIDIA CORPORATION. All rights reserved.
|
|
*
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*/
|
|
|
|
#pragma once
|
|
|
|
#include "tensorrt_llm/common/cudaUtils.h"
|
|
#include <assert.h>
|
|
#include <cuda_fp16.h>
|
|
#include <cuda_runtime.h>
|
|
|
|
namespace tensorrt_llm
|
|
{
|
|
namespace kernels
|
|
{
|
|
|
|
template <typename Tout>
|
|
void invokeRecoverFromRA(Tout* accu_output, float* accu_softmax_stats, Tout* output, float* softmax_stats, int b, int s,
|
|
int h, int d, int* cu_seqlens, cudaStream_t stream);
|
|
|
|
} // namespace kernels
|
|
} // namespace tensorrt_llm
|