espnet2.asr.frontend.cnn.ConvLayerBlock

source

class espnet2.asr.frontend.cnn.ConvLayerBlock(in_channels: int, out_channels: int, kernel_size: int, stride: int, bias: bool, layer_norm: Module | None, conv_mode: str)

Bases: Module

Convolution unit of FeatureExtractor

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

forward(x: Tensor, length: Tensor | None) → Tuple[Tensor, Tensor | None]

• Parameters:
  • x (Tensor) – Shape: [batch, in_channels, in_frame].
  • length (Tensor or None , optional) – Shape [batch, ].
• Returns: Shape [batch, out_channels, out_frames]. Optional[Tensor]: Shape [batch, ].
• Return type: Tensor