Calculate Global Layer Normalization dim: (int or list or torch.Size) input shape from an expected input of size e

Container module for a single RNN layer. args: rnn_type: string, select from 'RNN', 'LSTM' and 'GRU'. input_size: int,

r"""TransformerEncoderLayer is made up of self-attn and feedforward network. This standard encoder layer is based on the paper "Attention Is All

Deep duaL-path RNN. args: rnn_type: string, select from 'RNN', 'LSTM' and 'GRU'. input_size: int, dimension of the inpu

Deep duaL-path RNN. args: rnn_type: string, select from 'RNN', 'LSTM' and 'GRU'. input_size: int, dimension of the input

Calculate Cumulative Layer Normalization dim: you want to norm dim elementwise_affine: learnable per-element affine parameter