mmagic.models.editors.mspie.mspie_stylegan2_modules

Module Contents

Classes

ModulatedPEConv2d	Modulated Conv2d in StyleGANv2 with Positional Encoding (PE).
ModulatedPEStyleConv	Modulated Style Convolution with Positional Encoding.

Attributes

conv2d

mmagic.models.editors.mspie.mspie_stylegan2_modules.conv2d

class mmagic.models.editors.mspie.mspie_stylegan2_modules.ModulatedPEConv2d(in_channels, out_channels, kernel_size, style_channels, demodulate=True, upsample=False, downsample=False, blur_kernel=[1, 3, 3, 1], equalized_lr_cfg=dict(mode='fan_in', lr_mul=1.0, gain=1.0), style_mod_cfg=dict(bias_init=1.0), style_bias=0.0, eps=1e-08, no_pad=False, deconv2conv=False, interp_pad=None, up_config=dict(scale_factor=2, mode='nearest'), up_after_conv=False)

Bases: mmengine.model.BaseModule

Modulated Conv2d in StyleGANv2 with Positional Encoding (PE).

This module is modified from the ModulatedConv2d in StyleGAN2 to support the experiments in: Positional Encoding as Spatial Inductive Bias in GANs, CVPR’2021.

Parameters

• in_channels (int) – Input channels.

• out_channels (int) – Output channels.

• kernel_size (int) – Kernel size, same as nn.Con2d.

• style_channels (int) – Channels for the style codes.

• demodulate (bool, optional) – Whether to adopt demodulation. Defaults to True.

• upsample (bool, optional) – Whether to adopt upsampling in features. Defaults to False.

• downsample (bool, optional) – Whether to adopt downsampling in features. Defaults to False.

• blur_kernel (list[int], optional) – Blurry kernel. Defaults to [1, 3, 3, 1].

• equalized_lr_cfg (dict | None, optional) – Configs for equalized lr. Defaults to dict(mode=’fan_in’, lr_mul=1., gain=1.).

• style_mod_cfg (dict, optional) – Configs for style modulation module. Defaults to dict(bias_init=1.).

• style_bias (float, optional) – Bias value for style code. Defaults to 0..

• eps (float, optional) – Epsilon value to avoid computation error. Defaults to 1e-8.

• no_pad (bool, optional) – Whether to removing the padding in convolution. Defaults to False.

• deconv2conv (bool, optional) – Whether to substitute the transposed conv with (conv2d, upsampling). Defaults to False.

• interp_pad (int | None, optional) – The padding number of interpolation pad. Defaults to None.

• up_config (dict, optional) – Upsampling config. Defaults to dict(scale_factor=2, mode=’nearest’).

• up_after_conv (bool, optional) – Whether to adopt upsampling after convolution. Defaults to False.

forward(x, style)

Forward function.

Parameters

• x ([Tensor) – Input features with shape of (N, C, H, W).

• style (Tensor) – Style latent with shape of (N, C).

Returns

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

Return type

Tensor

class mmagic.models.editors.mspie.mspie_stylegan2_modules.ModulatedPEStyleConv(in_channels, out_channels, kernel_size, style_channels, upsample=False, blur_kernel=[1, 3, 3, 1], demodulate=True, style_mod_cfg=dict(bias_init=1.0), style_bias=0.0, **kwargs)

Bases: mmengine.model.BaseModule

Modulated Style Convolution with Positional Encoding.

This module is modified from the ModulatedStyleConv in StyleGAN2 to support the experiments in: Positional Encoding as Spatial Inductive Bias in GANs, CVPR’2021.

Parameters

• in_channels (int) – Input channels.

• out_channels (int) – Output channels.

• kernel_size (int) – Kernel size, same as nn.Con2d.

• style_channels (int) – Channels for the style codes.

• demodulate (bool, optional) – Whether to adopt demodulation. Defaults to True.

• upsample (bool, optional) – Whether to adopt upsampling in features. Defaults to False.

• downsample (bool, optional) – Whether to adopt downsampling in features. Defaults to False.

• blur_kernel (list[int], optional) – Blurry kernel. Defaults to [1, 3, 3, 1].

• equalized_lr_cfg (dict | None, optional) – Configs for equalized lr. Defaults to dict(mode=’fan_in’, lr_mul=1., gain=1.).

• style_mod_cfg (dict, optional) – Configs for style modulation module. Defaults to dict(bias_init=1.).

• style_bias (float, optional) – Bias value for style code. Defaults to 0..

forward(x, style, noise=None, return_noise=False)

Forward Function.

Parameters

• x ([Tensor) – Input features with shape of (N, C, H, W).

• style (Tensor) – Style latent with shape of (N, C).

• noise (Tensor, optional) – Noise for injection. Defaults to None.

• return_noise (bool, optional) – Whether to return noise tensors. Defaults to False.

Returns

Output features with shape of (N, C, H, W)

Return type

Tensor