espnet.nets.pytorch_backend.transformer.embedding.PositionalEncoding

source

class espnet.nets.pytorch_backend.transformer.embedding.PositionalEncoding(d_model, dropout_rate, max_len=5000, reverse=False)

Bases: Module

Positional encoding.

• Parameters:
  • d_model (int) – Embedding dimension.
  • dropout_rate (float) – Dropout rate.
  • max_len (int) – Maximum input length.
  • reverse (bool) – Whether to reverse the input position. Only for
  • current (the class LegacyRelPositionalEncoding. We remove it in the)
  • RelPositionalEncoding. (class)

Construct an PositionalEncoding object.

extend_pe(x)

Reset the positional encodings.

forward(x: Tensor)

Add positional encoding.

• Parameters:x (torch.Tensor) – Input tensor (batch, time, *).
• Returns: Encoded tensor (batch, time, *).
• Return type: torch.Tensor