DeconvolutionLayer.cpp

Go to the documentation of this file.

/*
 * Copyright (c) 2017-2021 Arm Limited.
 *
 * SPDX-License-Identifier: MIT
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to
 * deal in the Software without restriction, including without limitation the
 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
 * sell copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all
 * copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */
#include "ConvolutionLayer.h"
 
#include "tests/validation/Helpers.h"
 
namespace arm_compute
{
namespace test
{
namespace validation
{
namespace reference
{
template <typename T, typename TW, typename TB>
SimpleTensor<T> deconvolution_layer(const SimpleTensor<T> &src, const SimpleTensor<TW> &weights, const SimpleTensor<TB> &bias, const TensorShape &output_shape,
                                    const PadStrideInfo &info, QuantizationInfo out_qinfo)
{
    // Create reference
    const unsigned int pad_left           = info.pad_left();
    const unsigned int pad_right          = info.pad_right();
    const unsigned int pad_top            = info.pad_top();
    const unsigned int pad_bottom         = info.pad_bottom();
    const int          stride_x           = info.stride().first;
    const int          stride_y           = info.stride().second;
    const int          weights_width      = weights.shape().x();
    const int          weights_height     = weights.shape().y();
    const int          weights_upper_dims = weights.shape().total_size() / (weights_width * weights_height);
 
    ARM_COMPUTE_ERROR_ON(pad_left > (weights.shape().x() - 1));
    ARM_COMPUTE_ERROR_ON(pad_right > (weights.shape().x() - 1));
    ARM_COMPUTE_ERROR_ON(pad_top > (weights.shape().y() - 1));
    ARM_COMPUTE_ERROR_ON(pad_bottom > (weights.shape().y() - 1));
 
    // Find the upsampled dimensions
    unsigned int out_x = (src.shape().x() - 1) * stride_x + 1;
    unsigned int out_y = (src.shape().y() - 1) * stride_y + 1;
 
    // Find the padding needed for the convolution with stride 1 in order to match output shape
    unsigned int deconv_pad_x = output_shape.x() - (out_x - weights_width + 1);
    unsigned int deconv_pad_y = output_shape.y() - (out_y - weights_height + 1);
    out_x += deconv_pad_x;
    out_y += deconv_pad_y;
 
    unsigned int deconv_pad_left  = pad_right > pad_left ? pad_right - pad_left : 0;
    unsigned int deconv_pad_right = pad_left > pad_right ? pad_left - pad_right : 0;
    deconv_pad_x -= deconv_pad_left + deconv_pad_right;
    ARM_COMPUTE_ERROR_ON((deconv_pad_x % 2) != 0);
    deconv_pad_left += deconv_pad_x / 2;
    deconv_pad_right += deconv_pad_x / 2;
 
    unsigned int deconv_pad_top    = pad_bottom > pad_top ? pad_bottom - pad_top : 0;
    unsigned int deconv_pad_bottom = pad_top > pad_bottom ? pad_top - pad_bottom : 0;
    deconv_pad_y -= deconv_pad_top + deconv_pad_bottom;
    ARM_COMPUTE_ERROR_ON((deconv_pad_y % 2) != 0);
    deconv_pad_top += deconv_pad_y / 2;
    deconv_pad_bottom += deconv_pad_y / 2;
 
    TensorShape scaled_shape = src.shape();
    scaled_shape.set(0, out_x);
    scaled_shape.set(1, out_y);
    SimpleTensor<T> scaled{ scaled_shape, src.data_type(), 1, src.quantization_info() };
 
    const int width_in      = src.shape().x();
    const int height_in     = src.shape().y();
    const int width_scaled  = scaled.shape().x();
    const int height_scaled = scaled.shape().y();
    const int num_2d_slices = src.shape().total_size() / (width_in * height_in);
 
    if(src.data_type() == DataType::QASYMM8 || src.data_type() == DataType::QASYMM8_SIGNED)
    {
        const auto quantized_zero = static_cast<T>(src.quantization_info().uniform().offset);
        std::fill_n(scaled.data(), scaled.num_elements(), quantized_zero);
    }
    else
    {
        std::fill_n(scaled.data(), scaled.num_elements(), T(0));
    }
 
    // Flip weights by 180 degrees
    SimpleTensor<TW> weights_flipped{ weights.shape(), weights.data_type(), 1, weights.quantization_info(), weights.data_layout() };
#if defined(_OPENMP)
    #pragma omp parallel for
#endif /* _OPENMP */
    for(int ud = 0; ud < weights_upper_dims; ++ud)
    {
        const int offset = ud * weights_width * weights_height;
        for(int y = 0; y < weights_height; ++y)
        {
            for(int x = 0; x < weights_width; ++x)
            {
                weights_flipped[offset + (weights_height - 1 - y) * weights_width + (weights_width - 1 - x)] = weights[offset + y * weights_width + x];
            }
        }
    }
#if defined(_OPENMP)
    #pragma omp parallel for
#endif /* _OPENMP */
    for(int slice = 0; slice < num_2d_slices; ++slice)
    {
        const int offset_slice_in  = slice * width_in * height_in;
        const int offset_slice_out = slice * width_scaled * height_scaled;
        const int start_x          = deconv_pad_left;
        const int start_y          = deconv_pad_top;
        const int end_x            = width_scaled - deconv_pad_right;
        const int end_y            = height_scaled - deconv_pad_bottom;
 
        for(int yi = start_y, in_y = 0; yi < end_y; yi += stride_y, in_y++)
        {
            for(int xi = start_x, in_x = 0; xi < end_x; xi += stride_x, in_x++)
            {
                const T *in  = src.data() + offset_slice_in + in_y * width_in + in_x;
                T       *out = scaled.data() + offset_slice_out + xi + yi * width_scaled;
                *out         = *in;
            }
        }
    }
 
    const PadStrideInfo conv_info(1, 1, 0, 0, 0, 0, DimensionRoundingType::CEIL);
    return convolution_layer(scaled, weights_flipped, bias, output_shape, conv_info, Size2D(1U, 1U), 1, out_qinfo);
}
 
template SimpleTensor<uint8_t> deconvolution_layer(const SimpleTensor<uint8_t> &src, const SimpleTensor<uint8_t> &weights, const SimpleTensor<int32_t> &bias, const TensorShape &output_shape,
                                                   const PadStrideInfo &info, QuantizationInfo out_quant_info);
template SimpleTensor<uint8_t> deconvolution_layer(const SimpleTensor<uint8_t> &src, const SimpleTensor<int8_t> &weights, const SimpleTensor<int32_t> &bias, const TensorShape &output_shape,
                                                   const PadStrideInfo &info, QuantizationInfo out_quant_info);
template SimpleTensor<int8_t> deconvolution_layer(const SimpleTensor<int8_t> &src, const SimpleTensor<int8_t> &weights, const SimpleTensor<int32_t> &bias, const TensorShape &output_shape,
                                                  const PadStrideInfo &info, QuantizationInfo out_quant_info);
template SimpleTensor<float> deconvolution_layer(const SimpleTensor<float> &src, const SimpleTensor<float> &weights, const SimpleTensor<float> &bias, const TensorShape &output_shape,
                                                 const PadStrideInfo &info, QuantizationInfo out_quant_info);
template SimpleTensor<half> deconvolution_layer(const SimpleTensor<half> &src, const SimpleTensor<half> &weights, const SimpleTensor<half> &bias, const TensorShape &output_shape,
                                                const PadStrideInfo &info, QuantizationInfo out_quant_info);
} // namespace reference
} // namespace validation
} // namespace test
} // namespace arm_compute