e8b43356aa8cb8c659cae25aea32a49fd0881cd9,gluon/gluoncv2/models/densenet_cifar.py,CIFARDenseNet,__init__,#CIFARDenseNet#Any#Any#Any#Any#Any#Any#Any#,38

Before Change

        self.in_size = in_size
        self.classes = classes

        with self.name_scope():
            self.features = nn.HybridSequential(prefix="")
            self.features.add(conv3x3(
                in_channels=in_channels,
                out_channels=init_block_channels))
            in_channels = init_block_channels
            for i, channels_per_stage in enumerate(channels):
                stage = nn.HybridSequential(prefix="stage{}_".format(i + 1))
                with stage.name_scope():
                    if i != 0:
                        stage.add(TransitionBlock(
                            in_channels=in_channels,
                            out_channels=(in_channels // 2),
                            bn_use_global_stats=bn_use_global_stats))
                        in_channels = in_channels // 2
                    for j, out_channels in enumerate(channels_per_stage):
                        stage.add(DenseUnit(
                            in_channels=in_channels,
                            out_channels=out_channels,
                            bn_use_global_stats=bn_use_global_stats,
                            dropout_rate=dropout_rate))
                        in_channels = out_channels
                self.features.add(stage)
            self.features.add(PreResActivation(
                in_channels=in_channels,
                bn_use_global_stats=bn_use_global_stats))
            self.features.add(nn.AvgPool2D(
                pool_size=8,
                strides=1))

            self.output = nn.HybridSequential(prefix="")
            self.output.add(nn.Flatten())
            self.output.add(nn.Dense(
                units=classes,
                in_units=in_channels))

    def hybrid_forward(self, F, x):
        x = self.features(x)
        x = self.output(x)
        return x

After Change

    in_size : tuple of two ints, default (32, 32)
        Spatial size of the expected input image.
    classes : int, default 10
        Number of classification classes.
    
    def __init__(self,
                 channels,
                 init_block_channels,
                 bottleneck,
                 bn_use_global_stats=False,
                 dropout_rate=0.0,
                 in_channels=3,
                 in_size=(32, 32),
                 classes=10,
                 **kwargs):
        super(CIFARDenseNet, self).__init__(**kwargs)
        self.in_size = in_size
        self.classes = classes
        unit_class = DenseUnit if bottleneck else DenseSimpleUnit

        with self.name_scope():
            self.features = nn.HybridSequential(prefix="")
            self.features.add(conv3x3(
                in_channels=in_channels,
                out_channels=init_block_channels))
            in_channels = init_block_channels
            for i, channels_per_stage in enumerate(channels):
                stage = nn.HybridSequential(prefix="stage{}_".format(i + 1))
                with stage.name_scope():
                    if i != 0:
                        stage.add(TransitionBlock(
                            in_channels=in_channels,
                            out_channels=(in_channels // 2),
                            bn_use_global_stats=bn_use_global_stats))
                        in_channels = in_channels // 2
                    for j, out_channels in enumerate(channels_per_stage):
                        stage.add(unit_class(
                            in_channels=in_channels,

Instances