wehub-resource-sync
50 lines
1.9 KiB
C++
50 lines
1.9 KiB
C++
// Copyright (c) 2024 PaddlePaddle Authors. 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 "paddle/phi/core/dense_tensor.h"
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
|
|
namespace phi {
|
|
namespace fusion {
|
|
|
|
template <typename T, typename Context>
|
|
void MMHAKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& cache_kv,
|
|
const optional<DenseTensor>& bias,
|
|
const optional<DenseTensor>& src_mask,
|
|
const optional<DenseTensor>& cum_offsets,
|
|
const optional<DenseTensor>& sequence_lengths,
|
|
const optional<DenseTensor>& rotary_tensor,
|
|
const optional<DenseTensor>& beam_cache_offset,
|
|
const optional<DenseTensor>& qkv_out_scale,
|
|
const optional<DenseTensor>& out_shift,
|
|
const optional<DenseTensor>& out_smooth,
|
|
int seq_len,
|
|
int rotary_emb_dims,
|
|
const bool use_neox_rotary_style,
|
|
const std::string& compute_dtype,
|
|
const float out_scale,
|
|
const int quant_round_type,
|
|
const float quant_max_bound,
|
|
const float quant_min_bound,
|
|
DenseTensor* out,
|
|
DenseTensor* cache_kv_out,
|
|
DenseTensor* beam_cache_offset_out);
|
|
|
|
} // namespace fusion
|
|
} // namespace phi
|