130 lines
5.7 KiB
C++
130 lines
5.7 KiB
C++
/*******************************************************************************
|
|
* Copyright (c) 2015-2018 Skymind, Inc.
|
|
*
|
|
* This program and the accompanying materials are made available under the
|
|
* terms of the Apache License, Version 2.0 which is available at
|
|
* https://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.
|
|
*
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
******************************************************************************/
|
|
|
|
//
|
|
// Created by raver119 on 30.11.17.
|
|
//
|
|
|
|
#include <ops/declarable/helpers/col2im.h>
|
|
|
|
namespace nd4j {
|
|
namespace ops {
|
|
namespace helpers {
|
|
|
|
// [bS, iC, kH, kW, oH, oW] is de-convoluted to [bS, iC, iH, iW]
|
|
template <typename T>
|
|
void col2im_(nd4j::LaunchContext & context, const NDArray& input, NDArray& output, const int sH, const int sW, const int pH, const int pW, const int iH, const int iW, const int dH, const int dW) {
|
|
|
|
auto imBuff = output.bufferAsT<T>();
|
|
auto colBuff = input.bufferAsT<T>();
|
|
auto imShapeBuffer = output.getShapeInfo();
|
|
auto colShapeBuffer = input.getShapeInfo();
|
|
auto colShape = shape::shapeOf(colShapeBuffer);
|
|
auto colStride = shape::stride(colShapeBuffer);
|
|
auto imShape = shape::shapeOf(imShapeBuffer);
|
|
auto imStride = shape::stride(imShapeBuffer);
|
|
|
|
const int bS = imShape[0];
|
|
const int iC = imShape[1];
|
|
const int kH = colShape[2];
|
|
const int kW = colShape[3];
|
|
const int oH = colShape[4];
|
|
const int oW = colShape[5];
|
|
const Nd4jLong colStride0 = colStride[0];
|
|
const Nd4jLong colStride1 = colStride[1];
|
|
const Nd4jLong colStride2 = colStride[2];
|
|
const Nd4jLong colStride3 = colStride[3];
|
|
const Nd4jLong colStride4 = colStride[4];
|
|
const Nd4jLong colStride5 = colStride[5];
|
|
const Nd4jLong imStride0 = imStride[0];
|
|
const Nd4jLong imStride1 = imStride[1];
|
|
const Nd4jLong imStride2 = imStride[2];
|
|
const Nd4jLong imStride3 = imStride[3];
|
|
|
|
memset(imBuff, 0, shape::length(imShapeBuffer) * sizeof(T));
|
|
|
|
T *col, *im;
|
|
int imRow, imCol;
|
|
|
|
// if (shape::order(colShapeBuffer) == 'c' && shape::order(imShapeBuffer) == 'c' && shape::strideDescendingCAscendingF(colShapeBuffer) && shape::strideDescendingCAscendingF(imShapeBuffer)) {
|
|
if (false) {
|
|
|
|
PRAGMA_OMP_PARALLEL_FOR_SIMD_ARGS(private(col, im, imRow, imCol) collapse(2))
|
|
for (int b = 0; b < bS; b++) {
|
|
for (int c = 0; c < iC; ++c) {
|
|
for (int kRow = 0; kRow < kH; ++kRow) {
|
|
for (int kCol = 0; kCol < kW; ++kCol) {
|
|
for (int colH = 0; colH < oH; ++colH) {
|
|
for (int colW = 0; colW < oW; ++colW) {
|
|
|
|
imRow = (-pH + kRow * dH) + colH*sH;
|
|
imCol = (-pW + kCol * dW) + colW*sW;
|
|
|
|
col = colBuff + b*colStride0 + c*colStride1 + kRow*colStride2 + kCol*colStride3 + colH*colStride4 + colW*colStride5;
|
|
im = imBuff + b*imStride0 + c*imStride1 + imRow*imStride2 + imCol*imStride3;
|
|
|
|
if (static_cast<unsigned>(imRow) < static_cast<unsigned>(iH) && static_cast<unsigned>(imCol) < static_cast<unsigned>(iW))
|
|
*im += *col;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
|
|
PRAGMA_OMP_PARALLEL_FOR_SIMD_ARGS(firstprivate(imRow, imCol))
|
|
for (int b = 0; b < bS; ++b) {
|
|
T* im0 = imBuff + b*imStride0;
|
|
T* col4 = colBuff + b*colStride0;
|
|
for (int colH = 0; colH < oH; ++colH, col4 += colStride4) {
|
|
T* col5 = col4;
|
|
for (int colW = 0; colW < oW; ++colW, col5 += colStride5) {
|
|
T* col1 = col5;
|
|
T* im1 = im0;
|
|
for (int c = 0; c < iC; ++c, col1 += colStride1, im1 += imStride1) {
|
|
int imRow = (-pH + colH*sH);
|
|
T* col2 = col1;
|
|
T* im2 = im1 + imRow*imStride2;
|
|
for (int kRow = 0; kRow < kH; ++kRow, col2 += colStride2, imRow += dH, im2 += dH*imStride2) {
|
|
int imCol =-pW + colW*sW;
|
|
T* col3 = col2;
|
|
T* im3 = im2 + imCol*imStride3;
|
|
for (int kCol = 0; kCol < kW; ++kCol, col3 += colStride3, imCol += dW, im3 += dW*imStride3) {
|
|
|
|
if (static_cast<unsigned>(imRow) < static_cast<unsigned>(iH) && static_cast<unsigned>(imCol) < static_cast<unsigned>(iW))
|
|
*im3 += *col3;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
void col2im(nd4j::LaunchContext & context, const NDArray& input, NDArray& output, const int sH, const int sW, const int pH, const int pW, const int iH, const int iW, const int dH, const int dW) {
|
|
BUILD_SINGLE_SELECTOR(input.dataType(), col2im_, (context, input, output, sH, sW, pH, pW, iH, iW, dH, dW), LIBND4J_TYPES);
|
|
}
|
|
|
|
BUILD_SINGLE_TEMPLATE(template void col2im_, (nd4j::LaunchContext & context, const NDArray& input, NDArray& output, const int sH, const int sW, const int pH, const int pW, const int iH, const int iW, const int dH, const int dW), LIBND4J_TYPES);
|
|
|
|
}
|
|
}
|
|
}
|