src/runtime/CL/functions/CLFullyConnectedLayer.cpp

/*
 * Copyright (c) 2017 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 "arm_compute/runtime/CL/functions/CLFullyConnectedLayer.h"

#include "arm_compute/core/Validate.h"
#include "arm_compute/runtime/CL/CLScheduler.h"

#include <algorithm>
#include <cmath>

using namespace arm_compute;

CLFullyConnectedLayer::CLFullyConnectedLayer()
    : _im2col_kernel(), _transpose_kernel(), _transpose1xW_kernel(), _interleave4x4_kernel(), _mm_kernel(), _accumulate_biases_kernel(), _im2col_output(), _interleave4x4_output(), _transpose_output(),
      _transpose1xW_output(), _is_first_run(true), _transpose_weights(true), _fc_after_conv(true), _batched_fc_layer(false), _accumulate_biases(false)
{
}

void CLFullyConnectedLayer::configure_conv_fc_wb(const ICLTensor *input, const ICLTensor *weights, ICLTensor *output)
{
    ARM_COMPUTE_ERROR_ON(weights->info()->dimension(1) != (input->info()->dimension(0) * input->info()->dimension(1) * input->info()->dimension(2)));

    // If the fully connected layer is called after a convolution layer, the input tensor must be linearized

    // Initialize output tensor for im2col
    TensorShape shape_im2col;
    shape_im2col.set(0, weights->info()->dimension(1));
    shape_im2col.set(1, input->info()->dimension(3));
    shape_im2col.set(2, input->info()->dimension(4));
    shape_im2col.set(3, input->info()->dimension(5));
    _im2col_output.allocator()->init(TensorInfo(shape_im2col, 1, input->info()->data_type()));

    // Initialize output tensor for interleave 4x4
    TensorShape shape_interleaved = _im2col_output.info()->tensor_shape();
    shape_interleaved.set(0, shape_interleaved.x() * 4);
    shape_interleaved.set(1, std::ceil(static_cast<float>(shape_interleaved.y()) / 4));
    _interleave4x4_output.allocator()->init(TensorInfo(shape_interleaved, 1, input->info()->data_type()));

    // Initialize output tensor for transpose 1xW
    TensorShape shape_transposed1xW(weights->info()->dimension(1) * 4, static_cast<size_t>(std::ceil(weights->info()->dimension(0) / 4.f)));
    _transpose1xW_output.allocator()->init(TensorInfo(shape_transposed1xW, 1, weights->info()->data_type()));

    // Configure im2col kernel
    _im2col_kernel.configure(input, &_im2col_output, std::make_pair(1, 1), PadStrideInfo(1, 1, 0, 0), false);

    // Configure interleave4x4 kernel
    _interleave4x4_kernel.configure(&_im2col_output, &_interleave4x4_output);

    // Configure transpose 1xW kernel
    _transpose1xW_kernel.configure(weights, &_transpose1xW_output);

    // Configure matrix multiply kernel
    _mm_kernel.configure(&_interleave4x4_output, &_transpose1xW_output, output, 1.0f);

    // Allocate the tensors once all the configure methods have been called
    _im2col_output.allocator()->allocate();
    _interleave4x4_output.allocator()->allocate();
    _transpose1xW_output.allocator()->allocate();
}

void CLFullyConnectedLayer::configure_fc_fc_wb(const ICLTensor *input, const ICLTensor *weights, ICLTensor *output)
{
    // Initialize output tensor for interleave 4x4
    TensorShape shape_interleaved = input->info()->tensor_shape();
    shape_interleaved.set(0, shape_interleaved.x() * 4);
    shape_interleaved.set(1, std::ceil(static_cast<float>(shape_interleaved.y()) / 4));
    _interleave4x4_output.allocator()->init(TensorInfo(shape_interleaved, 1, input->info()->data_type()));

    // Initialize output tensor for transpose 1xW
    TensorShape shape_transposed1xW(weights->info()->dimension(1) * 4, static_cast<size_t>(std::ceil(weights->info()->dimension(0) / 4.f)));
    _transpose1xW_output.allocator()->init(TensorInfo(shape_transposed1xW, 1, weights->info()->data_type()));

    // Configure interleave4x4 kernel
    _interleave4x4_kernel.configure(input, &_interleave4x4_output);

    // Configure transpose 1xW kernel
    _transpose1xW_kernel.configure(weights, &_transpose1xW_output);

    // Configure matrix multiply kernel
    _mm_kernel.configure(&_interleave4x4_output, &_transpose1xW_output, output, 1.0f);

    // Allocate the tensors once all the configure methods have been called
    _interleave4x4_output.allocator()->allocate();
    _transpose1xW_output.allocator()->allocate();
}

void CLFullyConnectedLayer::configure_conv_fc_nb(const ICLTensor *input, const ICLTensor *weights, ICLTensor *output)
{
    ARM_COMPUTE_ERROR_ON((weights->info()->dimension(1) != (input->info()->dimension(0) * input->info()->dimension(1) * input->info()->dimension(2))));

    // If the fully connected layer is called after a convolution layer, the input tensor must be linearized

    // Initialize output tensor for im2col
    TensorShape shape_im2col;
    shape_im2col.set(0, weights->info()->dimension(1));
    shape_im2col.set(1, 1);
    _im2col_output.allocator()->init(TensorInfo(shape_im2col, 1, input->info()->data_type()));

    // Configure im2col kernel
    _im2col_kernel.configure(input, &_im2col_output, std::make_pair(1, 1), PadStrideInfo(1, 1, 0, 0), false);

    // Configure matrix multiply kernel
    _mm_kernel.configure(&_im2col_output, weights, output, 1.0f);

    // Allocate the output tensor for im2col once all the configure methods have been called
    _im2col_output.allocator()->allocate();
}

void CLFullyConnectedLayer::configure_fc_fc_nb(const ICLTensor *input, const ICLTensor *weights, ICLTensor *output)
{
    ARM_COMPUTE_ERROR_ON(input->info()->dimension(0) != weights->info()->dimension(1));

    // Configure matrix multiply kernel
    _mm_kernel.configure(input, weights, output, 1.0f);
}

void CLFullyConnectedLayer::configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, bool transpose_weights)
{
    ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::F32);
    ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(weights, 1, DataType::F32);
    ARM_COMPUTE_ERROR_ON_MISMATCHING_DATA_TYPES(input, weights, output);
    ARM_COMPUTE_ERROR_ON(weights->info()->num_dimensions() != 2);

    const ICLTensor *weights_to_use = weights;

    _is_first_run      = true;
    _transpose_weights = transpose_weights;
    _fc_after_conv     = true;
    _batched_fc_layer  = false;
    _accumulate_biases = false;

    if(biases != nullptr)
    {
        ARM_COMPUTE_ERROR_ON_MISMATCHING_DATA_TYPES(input, biases);

        _accumulate_biases = true;

        // Configure accumulate biases kernel
        _accumulate_biases_kernel.configure(output, biases);
    }

    // Check if we need to transpose the weights
    if(_transpose_weights)
    {
        // Initialize the output tensor for transpose
        TensorShape shape_transposed(weights->info()->dimension(1), weights->info()->dimension(0));
        _transpose_output.allocator()->init(TensorInfo(shape_transposed, 1, weights->info()->data_type()));
        _transpose_kernel.configure(weights, &_transpose_output);

        weights_to_use = &_transpose_output;
    }

    // With the Fully Connected layer we can have 4 different cases:
    //  1) Convolution layer -> Fully Connected layer without batches
    //  2) Fully Connected layer -> Fully Connected layer without batches
    //  3) Convolution layer -> Fully Connected layer with batches
    //  4) Fully Connected layer -> Fully Connected layer with batches

    // Check if we have a fully connected layer with batches
    _batched_fc_layer = (output->info()->dimension(1) > 1);

    if(_batched_fc_layer)
    {
        _fc_after_conv = (TensorShape::num_max_dimensions >= 4) && (std::equal(input->info()->tensor_shape().cbegin() + 3,
                                                                               input->info()->tensor_shape().cend(),
                                                                               output->info()->tensor_shape().cbegin() + 1));

        if(_fc_after_conv)
        {
            // Fully Connected layer after a Convolution Layer with batches
            configure_conv_fc_wb(input, weights_to_use, output);
        }
        else
        {
            // Fully Connected layer after a Fully Connected Layer with batches
            configure_fc_fc_wb(input, weights_to_use, output);
        }
    }
    else
    {
        _fc_after_conv = (weights_to_use->info()->dimension(1) == (input->info()->dimension(0) * input->info()->dimension(1) * input->info()->dimension(2)));

        if(_fc_after_conv)
        {
            // Fully Connected layer after a Convolution Layer without batches
            configure_conv_fc_nb(input, weights_to_use, output);
        }
        else
        {
            // Fully Connected layer after a Fully Connected Layer without batches
            configure_fc_fc_nb(input, weights_to_use, output);
        }
    }

    // Allocate the transpose tensor if the transpose_weights flag is true and once all the configure methods have been called
    if(_transpose_weights)
    {
        _transpose_output.allocator()->allocate();
    }
}

void CLFullyConnectedLayer::run()
{
    // The reshape of the weights happens only once
    if(_is_first_run)
    {
        _is_first_run = false;

        if(_transpose_weights)
        {
            CLScheduler::get().enqueue(_transpose_kernel);
        }

        if(_batched_fc_layer)
        {
            CLScheduler::get().enqueue(_transpose1xW_kernel);
        }
    }

    // Linearize input if it comes from a convolutional layer
    if(_fc_after_conv)
    {
        CLScheduler::get().enqueue(_im2col_kernel, false);
    }

    // Interleave input
    if(_batched_fc_layer)
    {
        CLScheduler::get().enqueue(_interleave4x4_kernel, false);
    }

    // Run matrix multiply
    CLScheduler::get().enqueue(_mm_kernel, !_accumulate_biases);

    // Accumulate biases if provided
    if(_accumulate_biases)
    {
        CLScheduler::get().enqueue(_accumulate_biases_kernel);
    }
}