mmagic.models.editors.singan.singan_modules

Module Contents

Classes

GeneratorBlock	Generator block used in SinGAN.
DiscriminatorBlock	Discriminator Block used in SinGAN.

class mmagic.models.editors.singan.singan_modules.GeneratorBlock(in_channels, out_channels, kernel_size, padding, num_layers, base_channels, min_feat_channels, out_act_cfg=dict(type='Tanh'), stride=1, allow_no_residual=False, **kwargs)

Bases: torch.nn.Module

Generator block used in SinGAN.

Parameters

• in_channels (int) – Input channels.

• out_channels (int) – Output channels.

• num_scales (int) – The number of scales/stages in generator. Note that this number is counted from zero, which is the same as the original paper.

• kernel_size (int, optional) – Kernel size, same as nn.Conv2d. Defaults to 3.

• padding (int, optional) – Padding for the convolutional layer, same as nn.Conv2d. Defaults to 0.

• num_layers (int, optional) – The number of convolutional layers in each generator block. Defaults to 5.

• base_channels (int, optional) – The basic channels for convolutional layers in the generator block. Defaults to 32.

• min_feat_channels (int, optional) – Minimum channels for the feature maps in the generator block. Defaults to 32.

• out_act_cfg (dict | None, optional) – Configs for output activation layer. Defaults to dict(type=’Tanh’).

• stride (int, optional) – Same as nn.Conv2d. Defaults to 1.

• allow_no_residual (bool, optional) – Whether to allow no residual link in this block. Defaults to False.

forward(x, prev)

Forward function.

Parameters

• x (Tensor) – Input feature map.

• prev (Tensor) – Previous feature map.

Returns

Output feature map with the shape of (N, C, H, W).

Return type

Tensor

init_weights(pretrained=None)

class mmagic.models.editors.singan.singan_modules.DiscriminatorBlock(in_channels, base_channels, min_feat_channels, kernel_size, padding, num_layers, norm_cfg=dict(type='BN'), act_cfg=dict(type='LeakyReLU', negative_slope=0.2), stride=1, **kwargs)

Bases: torch.nn.Module

Discriminator Block used in SinGAN.

Parameters

• in_channels (int) – Input channels.

• base_channels (int) – Base channels for this block.

• min_feat_channels (int) – The minimum channels for feature map.

• kernel_size (int) – Size of convolutional kernel, same as nn.Conv2d.

• padding (int) – Padding for convolutional layer, same as nn.Conv2d.

• num_layers (int) – The number of convolutional layers in this block.

• norm_cfg (dict | None, optional) – Config for the normalization layer. Defaults to dict(type=’BN’).

• act_cfg (dict | None, optional) – Config for the activation layer. Defaults to dict(type=’LeakyReLU’, negative_slope=0.2).

• stride (int, optional) – The stride for the convolutional layer, same as nn.Conv2d. Defaults to 1.

forward(x)

Forward function.

Parameters

x (Tensor) – Input feature map with shape of (N, C, H, W).

Returns

Output feature map.

Return type

Tensor

init_weights(pretrained=None)