AI-MXNet
view release on metacpan or search on metacpan
t/test_symbol.t view on Meta::CPAN
# now we can do full shape inference
my $state_shape = $out_shapes->[0];
($arg_shapes, $out_shapes, $aux_shapes) = $out->infer_shape(data=>[$num_sample, $num_dim], prevstate=>$state_shape);
@arg_shapes{ @{ $out->list_arguments } } = @{ $arg_shapes };
is_deeply($arg_shapes{data}, [$num_sample, $num_dim]);
is_deeply($arg_shapes{x2h_weight}, [$num_hidden, $num_dim]);
is_deeply($arg_shapes{h2h_weight}, [$num_hidden, $num_hidden]);
}
test_symbol_infer_shape();
sub test_symbol_infer_shape_var
{
#Test specifying shape information when constructing a variable
my $shape = [2, 3];
my $a = mx->symbol->Variable('a', shape=>$shape);
my $b = mx->symbol->Variable('b');
my $c = mx->symbol->elemwise_add($a, $b);
my ($arg_shapes, $out_shapes, $aux_shapes) = $c->infer_shape();
is_deeply($arg_shapes->[0], $shape);
is_deeply($arg_shapes->[1], $shape);
is_deeply($out_shapes->[0], $shape);
$shape = [5, 6];
($arg_shapes, $out_shapes, $aux_shapes) = $c->infer_shape(a=>$shape);
is_deeply($arg_shapes->[0], $shape);
is_deeply($arg_shapes->[1], $shape);
is_deeply($out_shapes->[0], $shape);
}
test_symbol_infer_shape_var();
sub check_symbol_consistency
{
my ($sym1, $sym2, $ctx) = @_;
is_deeply($sym1->list_arguments(), $sym2->list_arguments());
is_deeply($sym1->list_auxiliary_states(), $sym2->list_auxiliary_states());
is_deeply($sym1->list_outputs(), $sym2->list_outputs());
check_consistency(sym => [$sym1, $sym2], ctx_list => [$ctx, $ctx]);
}
sub test_load_000800
{
my ($data, $weight, $fc1, $act1);
{
local($mx::AttrScope) = mx->AttrScope(ctx_group=>'stage1');
$data = mx->symbol->Variable('data', lr_mult=>0.2);
$weight = mx->sym->Variable('fc1_weight', lr_mult=>1.2);
$fc1 = mx->symbol->FullyConnected(data => $data, weight=>$weight, name=>'fc1', num_hidden=>128, wd_mult=>0.3);
$act1 = mx->symbol->Activation(data => $fc1, name=>'relu1', act_type=>"relu");
}
my ($fc2, $act2, $fc3, $sym1);
{
local($mx::AttrScope) = mx->AttrScope(ctx_group=>'stage2');
$fc2 = mx->symbol->FullyConnected(data => $act1, name => 'fc2', num_hidden => 64, lr_mult=>0.01);
$act2 = mx->symbol->Activation(data => $fc2, name=>'relu2', act_type=>"relu");
$fc3 = mx->symbol->FullyConnected(data => $act2, name=>'fc3', num_hidden=>10);
$fc3 = mx->symbol->BatchNorm($fc3, name=>'batchnorm0');
$sym1 = mx->symbol->SoftmaxOutput(data => $fc3, name => 'softmax')
}
{ local $/ = undef; my $json = <DATA>; open(F, ">save_000800.json"); print F $json; close(F); };
my $sym2 = mx->sym->load('save_000800.json');
unlink 'save_000800.json';
my %attr1 = %{ $sym1->attr_dict };
my %attr2 = %{ $sym2->attr_dict };
while(my ($k, $v1) = each %attr1)
{
ok(exists $attr2{ $k });
my $v2 = $attr2{$k};
while(my ($kk, $vv1) = each %{ $v1 })
{
if($kk =~ /^__/ and $kk =~ /__$/)
{
ok(exists $v2->{$kk} and $v2->{$kk} eq $vv1);
}
}
}
check_symbol_consistency($sym1, $sym2,
{ctx => mx->cpu(0), group2ctx =>{stage1 => mx->cpu(1), stage2 => mx->cpu(2) }, shapes => { data => [1,200] }}
);
}
test_load_000800();
__DATA__
{
"nodes": [
{
"op": "null",
"param": {},
"name": "data",
"inputs": [],
"backward_source_id": -1,
"attr": {
"ctx_group": "stage1",
"lr_mult": "0.2"
}
},
{
"op": "null",
"param": {},
"name": "fc1_weight",
"inputs": [],
"backward_source_id": -1,
"attr": {
"ctx_group": "stage1",
"wd_mult": "0.3",
"weight_lr_mult": "1.2"
}
},
{
"op": "null",
"param": {},
"name": "fc1_bias",
"inputs": [],
"backward_source_id": -1,
"attr": {
"ctx_group": "stage1",
"wd_mult": "0.3",
( run in 0.752 second using v1.01-cache-2.11-cpan-5837b0d9d2c )