Files
paddlepaddle--paddle/paddle/phi/kernels/linear_v2_kernel.h
T
2026-07-13 12:40:42 +08:00

54 lines
1.9 KiB
C++

// Copyright (c) 2025 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/backends/all_context.h"
#include "paddle/phi/core/dense_tensor.h"
#include "paddle/phi/core/enforce.h"
namespace phi {
// we don't receive 2+d tensor as weight
inline std::tuple<int64_t, int64_t, int64_t> canonicalize_dims(
const DenseTensor& input,
const DenseTensor& weight,
const bool transpose_weight) {
const auto input_dims = input.dims();
const auto weight_dims = weight.dims();
// We assume weight to be [K, N] if not tranasposed, [N, K] if transposed, [K]
// if 1D
const int64_t N = weight_dims.size() < 2 ? 1 : weight_dims[!transpose_weight];
const int64_t K =
weight_dims.size() < 2 ? weight_dims[0] : weight_dims[transpose_weight];
int64_t M = input_dims.size() >= 2 ? input_dims[input_dims.size() - 2] : 1;
if (input_dims.size() > 2) {
// Accumulate the batch dims for input
for (int64_t i = 0; i < input_dims.size() - 2; ++i) {
M *= input_dims[i];
}
}
return {M, N, K};
}
template <typename T, typename Context>
void LinearV2Kernel(const Context& dev_ctx,
const DenseTensor& input,
const DenseTensor& weight,
const DenseTensor& bias,
const bool transpose_weight,
DenseTensor* out);
} // namespace phi