I have one question about the input of nn.Linear after nn.LSTM.
I have the two different cases about the input of nn.Linear after nn.LSTM using same dataset in time-series, one using output of nn.LSTM as the input of the following nn.Linear, and the other using the hidden state of nn.LSTM as the input of the following nn.Linear.
are there any reasons for these cases?
the case with the output of nn.LSTM as the input of the following nn.Linear,
class LSTM(nn.Module):
def init(self, input_size=1, hidden_layer_size=100, output_size=1):
super().init()
self.hidden_layer_size = hidden_layer_size
self.lstm = nn.LSTM(input_size, hidden_layer_size)
self.linear = nn.Linear(hidden_layer_size, output_size)
self.hidden_cell = (torch.zeros(1,1,self.hidden_layer_size),
torch.zeros(1,1,self.hidden_layer_size))
def forward(self, input_seq):
lstm_out, self.hidden_cell = self.lstm(input_seq.view(len(input_seq) ,1, -1), self.hidden_cell)
predictions = self.linear(lstm_out.view(len(input_seq), -1))
return predictions[-1]
the case with the hidden state of nn.LSTM as the input of the following nn.Linear.
class LSTM(nn.Module):
def __init__(self, num_classes, input_size, hidden_size, num_layers):
super(LSTM, self).__init__()
self.num_classes = num_classes
self.num_layers = num_layers
self.input_size = input_size
self.hidden_size = hidden_size
self.seq_length = seq_length
self.lstm = nn.LSTM(input_size=input_size, hidden_size=hidden_size,
num_layers=num_layers, batch_first=True)
self.fc = nn.Linear(hidden_size, num_classes)
def forward(self, x):
h_0 = Variable(torch.zeros(
self.num_layers, x.size(0), self.hidden_size))
c_0 = Variable(torch.zeros(
self.num_layers, x.size(0), self.hidden_size))
# Propagate input through LSTM
ula, (h_out, _) = self.lstm(x, (h_0, c_0))
h_out = h_out.view(-1, self.hidden_size)
out = self.fc(h_out)
return out