Next, let's continue with implementing the encoder.
We'll start with the constructor:
class EncoderRNN(torch.nn.Module): def __init__(self, input_size, hidden_size): super(EncoderRNN, self).__init__() self.input_size = input_size self.hidden_size = hidden_size # Embedding for the input words self.embedding = torch.nn.Embedding(input_size, hidden_size) # The actual rnn sell self.rnn_cell = torch.nn.GRU(hidden_size, hidden_size)
The entry point is the self.embedding module. It will take the index of each word and it will return its assigned embedding vector. We will not use pretrained word vectors (such as GloVe), but nevertheless, the concept of embedding vectors is the same—it's just that we'll initialize them ...
