espnet.nets.pytorch_backend.rnn.attentions.initial_att

source

espnet.nets.pytorch_backend.rnn.attentions.initial_att(atype, eprojs, dunits, aheads, adim, awin, aconv_chans, aconv_filts, han_mode=False)

Instantiates a single attention module

• Parameters:
  • atype (str) – attention type
  • eprojs (int) – # projection-units of encoder
  • dunits (int) – # units of decoder
  • aheads (int) – # heads of multi head attention
  • adim (int) – attention dimension
  • awin (int) – attention window size
  • aconv_chans (int) – # channels of attention convolution
  • aconv_filts (int) – filter size of attention convolution
  • han_mode (bool) – flag to swith on mode of hierarchical attention
• Returns: The attention module