99 lines
3.2 KiB
C++
99 lines
3.2 KiB
C++
// Copyright (c) 2023 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/kernels/funcs/fc_functor.h"
|
|
|
|
namespace phi {
|
|
|
|
template <typename T, typename Context>
|
|
void QuantLinearKernel(const Context& dev_ctx,
|
|
const DenseTensor& x,
|
|
const DenseTensor& w,
|
|
const optional<DenseTensor>& bias,
|
|
int in_num_col_dims,
|
|
const std::string& activation_type,
|
|
bool padding_weights,
|
|
float scale_in,
|
|
const std::vector<float>& scale_weights,
|
|
int quant_round_type,
|
|
float quant_max_bound,
|
|
float quant_min_bound,
|
|
DenseTensor* y) {
|
|
bool with_relu = activation_type == "relu" ? true : false;
|
|
auto w_dims = w.dims();
|
|
|
|
auto input_dims = x.dims();
|
|
std::vector<int64_t> output_dims;
|
|
auto in_mat_dims = common::flatten_to_2d(input_dims, in_num_col_dims);
|
|
auto w_dims0 = padding_weights ? w_dims[0] - 4 : w_dims[0];
|
|
auto w_dims1 = padding_weights ? w_dims[1] - 4 : w_dims[1];
|
|
PADDLE_ENFORCE_EQ(
|
|
in_mat_dims[1],
|
|
w_dims0,
|
|
common::errors::InvalidArgument(
|
|
"The input's second dimension and weight's first dimension is "
|
|
"expected to be the same. But received input's second dimension is"
|
|
"%d, input's shape is %s; weight's first dimension is %d, weight's"
|
|
" shape is %s.",
|
|
in_mat_dims[1],
|
|
in_mat_dims,
|
|
w_dims0,
|
|
make_ddim({w_dims0, w_dims1})));
|
|
|
|
output_dims.reserve(static_cast<size_t>(in_num_col_dims + 1));
|
|
for (int i = 0; i < in_num_col_dims; ++i) {
|
|
output_dims.push_back(input_dims[i]);
|
|
}
|
|
output_dims.push_back(w_dims1);
|
|
|
|
y->Resize(output_dims);
|
|
y->set_lod(x.lod());
|
|
|
|
auto out_dims = y->dims();
|
|
int M = common::product(out_dims) / w_dims1;
|
|
|
|
const T* input_data = x.data<T>();
|
|
auto* output_data = dev_ctx.template Alloc<T>(y, y->numel() * sizeof(T));
|
|
auto bias_data = bias ? bias.get_ptr()->data<T>() : NULL;
|
|
|
|
PADDLE_ENFORCE_EQ(
|
|
w.dtype(),
|
|
DataType::INT8,
|
|
common::errors::InvalidArgument(
|
|
"The weight's datatype is expected to be int8 when use quant. But "
|
|
"received weight's datatype is %d",
|
|
static_cast<int>(w.dtype())));
|
|
funcs::FCInt8Functor<Context, T> fc;
|
|
fc(dev_ctx,
|
|
M,
|
|
w_dims1,
|
|
w_dims0,
|
|
input_data,
|
|
&w,
|
|
output_data,
|
|
scale_in,
|
|
scale_weights,
|
|
quant_round_type,
|
|
quant_max_bound,
|
|
quant_min_bound,
|
|
bias_data,
|
|
with_relu,
|
|
padding_weights);
|
|
return;
|
|
}
|
|
|
|
} // namespace phi
|