espnet2.gan_tts.wavenet.residual_block.Conv1d1x1

source

class espnet2.gan_tts.wavenet.residual_block.Conv1d1x1(in_channels: int, out_channels: int, bias: bool)

Bases: Conv1d

1x1 Conv1d with customized initialization.

This class implements a 1x1 convolutional layer that extends the Conv1d class with a specific initialization method for its weights.

in_channels

The number of input channels.

• Type: int

out_channels

The number of output channels.

• Type: int

bias

Whether to include a bias term in the convolution.

• Type: bool

• Parameters:

  • in_channels (int) – Number of input channels.
  • out_channels (int) – Number of output channels.
  • bias (bool) – Whether to add bias parameter in the convolution.

Examples

>>> conv1d_1x1 = Conv1d1x1(in_channels=16, out_channels=32, bias=True)
>>> input_tensor = torch.randn(1, 16, 10)  # (batch_size, channels, length)
>>> output_tensor = conv1d_1x1(input_tensor)
>>> output_tensor.shape
torch.Size([1, 32, 10])

NOTE

This layer is primarily used in the context of neural network architectures where 1x1 convolutions are beneficial, such as in residual networks.

Initialize 1x1 Conv1d module.