view release on metacpan or  search on metacpan

lib/AI/MXNet/Gluon/ModelZoo/Vision/MobileNet.pm  view on Meta::CPAN

# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements.  See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership.  The ASF licenses this file
# to you 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.

use strict;
use warnings;
use AI::MXNet::Function::Parameters;
package AI::MXNet::Gluon::ModelZoo::Vision::MobileNet::RELU6;
use AI::MXNet::Gluon::Mouse;
extends 'AI::MXNet::Gluon::HybridBlock';

method hybrid_forward(GluonClass $F, GluonInput $x)
{
    return $F->clip($x, a_min => 0, a_max => 6, name=>"relu6");
}

package AI::MXNet::Gluon::ModelZoo::Vision::MobileNet::LinearBottleneck;
use AI::MXNet::Gluon::Mouse;
extends 'AI::MXNet::Gluon::HybridBlock';
has [qw/in_channels channels t stride/] => (is => 'ro', isa => 'Int', required => 1);
method python_constructor_arguments(){ [qw/in_channels channels t stride/] }

=head1 NAME

    AI::MXNet::Gluon::ModelZoo::Vision::MobileNet::LinearBottleneck - LinearBottleneck used in MobileNetV2 model
=cut

=head1 DESCRIPTION

    LinearBottleneck used in MobileNetV2 model from the
    "Inverted Residuals and Linear Bottlenecks:
      Mobile Networks for Classification, Detection and Segmentation"
    <https://arxiv.org/abs/1801.04381> paper.

    Parameters
    ----------
    in_channels : Int
        Number of input channels.
    channels : Int
        Number of output channels.
    t : Int
        Layer expansion ratio.
    stride : Int
        stride
=cut

func _add_conv(
    $out, $channels, :$kernel=1, :$stride=1, :$pad=0,
    :$num_group=1, :$active=1, :$relu6=0
)
{
    $out->add(nn->Conv2D($channels, $kernel, $stride, $pad, groups=>$num_group, use_bias=>0));
    $out->add(nn->BatchNorm(scale=>1));
    if($active)
    {
        $out->add($relu6 ? AI::MXNet::Gluon::ModelZoo::Vision::MobileNet::RELU6->new : nn->Activation('relu'));
    }
}

sub BUILD
{
    my $self = shift;
    $self->use_shortcut($self->stride == 1 and $self->in_channels == $self->channels);
    $self->name_scope(sub {
        $self->out(nn->HybridSequential());
        _add_conv($self->out, $self->in_channels * $self->t, relu6=>1);
        _add_conv(
            $self->out, $self->in_channels * $self->t, kernel=>3, stride=>$self->stride,
            pad=>1, num_group=>$self->in_channels * $self->t, relu6=>1
        );
        _add_conv($self->out, $self->channels, active=>0, relu6=>1);
    });
}

method hybrid_forward($F, $x)
{
    my $out = $self->out->($x);
    if($self->use_shortcut)
    {
        $out = $F->elemwise_add($out, $x);