espnet2.spk.encoder.ska_tdnn_encoder.ResBlock

source

class espnet2.spk.encoder.ska_tdnn_encoder.ResBlock(inplanes: int, planes: int, stride: int = 1, reduction: int = 8, skfwse_freq: int = 40, skcwse_channel: int = 128)

Bases: Module

Residual Block with Selective Kernel Attention.

This class implements a residual block that incorporates selective kernel attention mechanisms for enhancing feature representation in deep learning models. It consists of a convolutional layer followed by batch normalization, ReLU activation, and two types of selective kernel attention: forward and channel-wise.

• Parameters:
  • inplanes (int) – Number of input channels.
  • planes (int) – Number of output channels.
  • stride (int , optional) – Stride of the convolution. Default is 1.
  • reduction (int , optional) – Reduction ratio for the attention mechanism. Default is 8.
  • skfwse_freq (int , optional) – Frequency parameter for forward selective kernel attention. Default is 40.
  • skcwse_channel (int , optional) – Number of channels for channel-wise selective kernel attention. Default is 128.

conv1

Convolutional layer.

• Type: nn.Conv2d

bn1

Batch normalization layer.

• Type: nn.BatchNorm2d

relu

ReLU activation function.

• Type: nn.ReLU

skfwse

Forward selective kernel attention module.

• Type:fwSKAttention

skcwse

Channel-wise selective kernel attention module.

• Type:cwSKAttention

stride

Stride of the convolution.

• Type: int

• Returns: The output of the forward pass.

• Return type: Tensor

####### Examples

>>> block = ResBlock(inplanes=64, planes=128)
>>> input_tensor = torch.randn(1, 64, 32, 32)  # Batch size 1, 64 channels
>>> output_tensor = block(input_tensor)
>>> print(output_tensor.shape)
torch.Size([1, 128, 32, 32])

Initialize internal Module state, shared by both nn.Module and ScriptModule.

forward(x)

Forward function for the ResBlock module.

This method takes an input tensor x, applies a series of convolutional, batch normalization, and activation operations, and then adds a residual connection from the input to the output.

• Parameters:x (torch.Tensor) – Input tensor of shape (B, C, H, W), where B is the batch size, C is the number of channels, H is the height, and W is the width.
• Returns: Output tensor of the same shape as input x.
• Return type: torch.Tensor

####### Examples

>>> res_block = ResBlock(inplanes=64, planes=128)
>>> input_tensor = torch.randn(32, 64, 16, 16)  # Batch of 32
>>> output_tensor = res_block(input_tensor)
>>> output_tensor.shape
torch.Size([32, 128, 16, 16])