def __init__(self,
input_size,
context_size,
hidden_size,
num_layers=1,
bias=True,
batch_first=False,
dropout=0):
"""
Constructor for the RecurrentAttention.
:param input_size: number of features in input tensor
:param context_size: number of features in output from encoder
:param hidden_size: internal hidden size
:param num_layers: number of layers in LSTM
:param bias: enables bias in LSTM layers
:param batch_first: if True the model uses (batch,seq,feature) tensors,
if false the model uses (seq, batch, feature)
:param dropout: probability of dropout
"""
super(RecurrentAttention, self).__init__()
self.rnn = nn.LSTM(input_size, hidden_size, num_layers, bias,
batch_first)
self.attn = BahdanauAttention(hidden_size,
context_size,
context_size,
normalize=True,
batch_first=batch_first)
self.dropout = nn.Dropout(dropout)
def __init__(self, input_size, context_size, hidden_size, num_layers=1,
bias=True, batch_first=False, dropout=0):
super(RecurrentAttention, self).__init__()
self.rnn = nn.LSTM(input_size, hidden_size, num_layers, bias,
batch_first)
# SSY seq2seq/models/attention.py
self.attn = BahdanauAttention(hidden_size, context_size, context_size,
normalize=True, batch_first=batch_first)
self.dropout = nn.Dropout(dropout)
def __init__(self,
input_size=1024,
context_size=1024,
hidden_size=1024,
num_layers=1,
batch_first=False,
dropout=0.2,
init_weight=0.1,
fusion=True):
"""
Constructor for the RecurrentAttention.
:param input_size: number of features in input tensor
:param context_size: number of features in output from encoder
:param hidden_size: internal hidden size
:param num_layers: number of layers in LSTM
:param batch_first: if True the model uses (batch,seq,feature) tensors,
if false the model uses (seq, batch, feature)
:param dropout: probability of dropout (on input to LSTM layer)
:param init_weight: range for the uniform initializer
"""
super(RecurrentAttention, self).__init__()
self.rnn = nn.LSTM(input_size,
hidden_size,
num_layers,
bias=True,
batch_first=batch_first)
init_lstm_(self.rnn, init_weight)
self.attn = BahdanauAttention(hidden_size,
context_size,
context_size,
normalize=True,
batch_first=batch_first,
fusion=fusion)
self.dropout = nn.Dropout(dropout)