AI-MXNet

 view release on metacpan or  search on metacpan

lib/AI/MXNet/Module.pm  view on Meta::CPAN

## TODO
## this class is here because of https://github.com/gfx/p5-Mouse/pull/67
## once 2.4.7 version of Mouse in Ubuntu for affected Perl version
## these accessors should be merged into main class

package AI::MXNet::Module::Private;
use Mouse;
has [qw/_param_names _fixed_param_names
        _aux_names _data_names _label_names _state_names
        _output_names _arg_params _aux_params
        _params_dirty _optimizer _kvstore
         _update_on_kvstore _updater _work_load_list
        _preload_opt_states _exec_group
        _data_shapes _label_shapes _context _grad_req/
] => (is => 'rw', init_arg => undef);

package AI::MXNet::Module;
use AI::MXNet::Base;
use AI::MXNet::Function::Parameters;
use List::Util qw(max);
use Data::Dumper ();
use Mouse;

func _create_kvstore(
    Maybe[Str|AI::MXNet::KVStore] $kvstore,
    Int                           $num_device,
    HashRef[AI::MXNet::NDArray]   $arg_params
)
{
    my $update_on_kvstore = 1;
    my $kv;
    if(defined $kvstore)
    {
        if(blessed $kvstore)
        {
            $kv = $kvstore;
        }
        else
        {
            # create kvstore using the string type
            if($num_device == 1 and $kvstore !~ /dist/)
            {
                # no need to use kv for single device and single machine
            }
            else
            {
                $kv = AI::MXNet::KVStore->create($kvstore);
                if($kvstore eq 'local')
                {
                    # automatically select a proper local
                    my $max_size = max(map { product(@{ $_->shape }) } values %{ $arg_params });
                    if($max_size > 1024 * 1024 * 16)
                    {
                        $update_on_kvstore = 0;
                    }
                }
            }
        }
    }

    $update_on_kvstore = 0 if not $kv;
    return ($kv, $update_on_kvstore);
}

func _initialize_kvstore(
    AI::MXNet::KVStore           :$kvstore,
    HashRef[AI::MXNet::NDArray]  :$arg_params,
    ArrayRef[Str]                :$param_names,
    Bool                         :$update_on_kvstore,
    ArrayRef[AI::MXNet::NDArray]|ArrayRef[ArrayRef[AI::MXNet::NDArray]] :$param_arrays
)
{
    enumerate(sub{
        my ($idx, $param_on_devs) = @_;
        my $name = $param_names->[$idx];
        $kvstore->init($name, $arg_params->{ $name });
        if($update_on_kvstore)
        {
            $kvstore->pull($name, out => $param_on_devs, priority => -$idx);
        }
    }, $param_arrays);
}

func _update_params_on_kvstore(
    ArrayRef[AI::MXNet::NDArray]|ArrayRef[ArrayRef[AI::MXNet::NDArray]] $param_arrays,
    ArrayRef[AI::MXNet::NDArray]|ArrayRef[ArrayRef[AI::MXNet::NDArray]] $grad_arrays,
    AI::MXNet::KVStore           $kvstore,
    ArrayRef[Str]                $param_names
)
{
    enumerate(sub{
        my ($index, $arg_list, $grad_list) = @_;
        if(ref $grad_list eq 'ARRAY' and not defined $grad_list->[0])
        {

lib/AI/MXNet/Module.pm  view on Meta::CPAN

method load_checkpoint(Str $prefix, Int $epoch)
{
    my $symbol = AI::MXNet::Symbol->load("$prefix-symbol.json");
    my %save_dict = %{ AI::MXNet::NDArray->load(sprintf('%s-%04d.params', $prefix, $epoch)) };
    my %arg_params;
    my %aux_params;
    while(my ($k, $v) = each %save_dict)
    {
        my ($tp, $name) = split(/:/, $k, 2);
        if($tp eq 'arg')
        {
            $arg_params{$name} = $v;
        }
        if($tp eq 'aux')
        {
            $aux_params{$name} = $v;
        }
    }
    return ($symbol, \%arg_params, \%aux_params);
}

=head1 NAME

    AI::MXNet::Module - FeedForward interface of MXNet.
    See AI::MXNet::Module::Base for the details.
=cut

extends 'AI::MXNet::Module::Base';

has '_symbol'           => (is => 'ro', init_arg => 'symbol', isa => 'AI::MXNet::Symbol', required => 1);
has '_data_names'       => (is => 'ro', init_arg => 'data_names', isa => 'ArrayRef[Str]');
has '_label_names'      => (is => 'ro', init_arg => 'label_names', isa => 'Maybe[ArrayRef[Str]]');
has 'work_load_list'    => (is => 'rw', isa => 'Maybe[ArrayRef[Int]]');
has 'fixed_param_names' => (is => 'rw', isa => 'Maybe[ArrayRef[Str]]');
has 'state_names'       => (is => 'rw', isa => 'Maybe[ArrayRef[Str]]');
has 'logger'            => (is => 'ro', default => sub { AI::MXNet::Logging->get_logger });
has '_p'                => (is => 'rw', init_arg => undef);
has 'context'           => (
    is => 'ro', 
    isa => 'AI::MXNet::Context|ArrayRef[AI::MXNet::Context]',
    default => sub { AI::MXNet::Context->cpu }
);

around BUILDARGS => sub {
    my $orig  = shift;
    my $class = shift;
    if(@_%2)
    {
        my $symbol = shift;
        return $class->$orig(symbol => $symbol, @_);
    }
    return $class->$orig(@_);
};

sub BUILD
{
    my $self = shift;
    $self->_p(AI::MXNet::Module::Private->new);
    my $context = $self->context;
    if(blessed $context)
    {
        $context = [$context];
    }
    $self->_p->_context($context);
    my $work_load_list = $self->work_load_list;
    if(not defined $work_load_list)
    {
        $work_load_list = [(1)x@{$self->_p->_context}];
    }
    assert(@{ $work_load_list } == @{ $self->_p->_context });
    $self->_p->_work_load_list($work_load_list);
    my @data_names  = @{ $self->_data_names//['data'] };
    my @label_names = @{ $self->_label_names//['softmax_label'] };
    my @state_names = @{ $self->state_names//[] };
    my $arg_names   = $self->_symbol->list_arguments;
    my @input_names = (@data_names, @label_names, @state_names);
    my %input_names = map { $_ => 1 } @input_names;
    $self->_p->_param_names([grep { not exists $input_names{$_} } @{ $arg_names }]);
    $self->_p->_fixed_param_names($self->fixed_param_names//[]);
    $self->_p->_state_names(\@state_names);
    $self->_p->_aux_names($self->_symbol->list_auxiliary_states);
    $self->_p->_data_names(\@data_names);
    $self->_p->_label_names(\@label_names);
    $self->_p->_output_names($self->_symbol->list_outputs);
    $self->_p->_params_dirty(0);
    $self->_check_input_names($self->_symbol, $self->_p->_data_names, "data", 1);
    $self->_check_input_names($self->_symbol, $self->_p->_label_names, "label", 0);
    $self->_check_input_names($self->_symbol, $self->_p->_state_names, "state", 1);
    $self->_check_input_names($self->_symbol, $self->_p->_fixed_param_names, "fixed_param", 1);
}

method Module(@args) { return @args ?  __PACKAGE__->new(@args) : __PACKAGE__ }
method BucketingModule(@args) { return AI::MXNet::Module::Bucketing->new(@args) }

=head2 load

        Create a model from previously saved checkpoint.

        Parameters
        ----------
        prefix : str
            path prefix of saved model files. You should have
            "prefix-symbol.json", "prefix-xxxx.params", and
            optionally "prefix-xxxx.states", where xxxx is the
            epoch number.
        epoch : int
            epoch to load.
        load_optimizer_states : bool
            whether to load optimizer states. Checkpoint needs
            to have been made with save_optimizer_states=True.
        data_names : array ref of str
            Default is ['data'] for a typical model used in image classification.
        label_names : array ref of str
            Default is ['softmax_label'] for a typical model used in image
            classification.
        logger : Logger
            Default is AI::MXNet::Logging.
        context : Context or list of Context
            Default is cpu(0).
        work_load_list : array ref of number

lib/AI/MXNet/Module.pm  view on Meta::CPAN

    {
        $self->borrow_optimizer($shared_module)
    }
}

=head2 reshape

    Reshape the module for new input shapes.
    Parameters
    ----------
    :$data_shapes : ArrayRef[AI::MXNet::DataDesc]
        Typically is $data_iter->provide_data.
    :$label_shapes= : Maybe[ArrayRef[AI::MXNet::DataDesc]]
        Typically is $data_iter->provide_label.
=cut

method reshape(
    ArrayRef[AI::MXNet::DataDesc|NameShape]        :$data_shapes,
    Maybe[ArrayRef[AI::MXNet::DataDesc|NameShape]] :$label_shapes=
)
{
    assert($self->binded);
    ($data_shapes, $label_shapes) = $self->_parse_data_desc(
        $self->data_names, $self->label_names, $data_shapes, $label_shapes
    );
    $self->_p->_data_shapes($data_shapes);
    $self->_p->_label_shapes($label_shapes);
    $self->_p->_exec_group->reshape($self->_p->_data_shapes, $self->_p->_label_shapes);
}

method init_optimizer(
    Str|AI::MXNet::KVStore :$kvstore='local',
    Optimizer              :$optimizer='sgd',
    HashRef                :$optimizer_params={ learning_rate => 0.01 },
    Bool                   :$force_init=0
)
{
    assert($self->binded and $self->params_initialized);
    if($self->optimizer_initialized and not $force_init)
    {
        $self->logger->warning('optimizer already initialized, ignoring...');
        return;
    }
    if($self->_p->_params_dirty)
    {
        $self->_sync_params_from_devices;
    }

    my ($kvstore, $update_on_kvstore) = _create_kvstore(
        $kvstore,
        scalar(@{$self->_p->_context}),
        $self->_p->_arg_params
    );
    my $batch_size = $self->_p->_exec_group->_p->batch_size;
    if($kvstore and $kvstore->type =~ /dist/ and $kvstore->type =~ /_sync/)
    {
        $batch_size *= $kvstore->num_workers;
    }
    my $rescale_grad = 1/$batch_size;

    if(not blessed $optimizer)
    {
        my %idx2name;
        if($update_on_kvstore)
        {
            @idx2name{ 0..@{$self->_p->_exec_group->param_names}-1 } = @{$self->_p->_exec_group->param_names};
        }
        else
        {
            for my $k (0..@{$self->_p->_context}-1)
            {
                @idx2name{ map { $_ + $k } 0..@{$self->_p->_exec_group->param_names}-1 } = @{$self->_p->_exec_group->param_names};
            }
        }
        if(not exists $optimizer_params->{rescale_grad})
        {
            $optimizer_params->{rescale_grad} = $rescale_grad;
        }
        $optimizer = AI::MXNet::Optimizer->create(
            $optimizer,
            sym  => $self->symbol,
            param_idx2name => \%idx2name,
            %{ $optimizer_params }
        );
        if($optimizer->rescale_grad != $rescale_grad)
        {
            AI::MXNet::Logging->warning(
                "Optimizer created manually outside Module but rescale_grad "
                ."is not normalized to 1.0/batch_size/num_workers (%s vs. %s). "
                ."Is this intended?",
                $optimizer->rescale_grad, $rescale_grad
            );
        }
    }

    $self->_p->_optimizer($optimizer);
    $self->_p->_kvstore($kvstore);
    $self->_p->_update_on_kvstore($update_on_kvstore);
    $self->_p->_updater(undef);

    if($kvstore)
    {
        # copy initialized local parameters to kvstore
        _initialize_kvstore(
            kvstore           => $kvstore,
            param_arrays      => $self->_p->_exec_group->_p->param_arrays,
            arg_params        => $self->_p->_arg_params,
            param_names       => $self->_p->_param_names,
            update_on_kvstore => $update_on_kvstore
        );
    }
    if($update_on_kvstore)
    {
        $kvstore->set_optimizer($self->_p->_optimizer);
    }
    else
    {
        $self->_p->_updater(AI::MXNet::Optimizer->get_updater($optimizer));
    }
    $self->optimizer_initialized(1);