input : visual feature [batch_size x T x input_size] output : contextual feature [batch_size x T x output_size]
(self, input)
| 69 | self.linear = nn.Linear(hidden_size * 2, output_size) |
| 70 | |
| 71 | def forward(self, input): |
| 72 | """ |
| 73 | input : visual feature [batch_size x T x input_size] |
| 74 | output : contextual feature [batch_size x T x output_size] |
| 75 | """ |
| 76 | self.rnn.flatten_parameters() |
| 77 | recurrent, _ = self.rnn(input) # batch_size x T x input_size -> batch_size x T x (2*hidden_size) |
| 78 | output = self.linear(recurrent) # batch_size x T x output_size |
| 79 | return output |
| 80 | |
| 81 | class BidirectionalRNN(nn.Module): |
| 82 |
nothing calls this directly
no outgoing calls
no test coverage detected