I am learning through PyTorch's seq2seq tutorial: https://pytorch.org/tutorials/intermediate/seq2seq_translation_tutorial.html I have a question about the decoder class DecoderRNN(nn.Module): def __init__(self, hidden_size, output_size): super(DecoderRNN, self).__init__() self.embedding = nn.Embedding(output_size, hidden_size) self.gru = nn.GRU(hidden_size, hidden_size, batch_first=True) self.out = nn.Linear(hidden_size, output_size) def forward(self, encoder_outputs, encoder_hidden, target_tensor=None): batch_size = encoder_outputs.size(0) decoder_input = torch.empty(batch_size, 1, dtype=torch.long, device=device).fill_(SOS_token) decoder_hidden = encoder_hidden decoder_outputs = [] for i in range(MAX_LENGTH): decoder_output, decoder_hidden = self.forward_step(decoder_input, decoder_hidden) decoder_outputs.append(decoder_output) if target_tensor is not None: # Teacher forcing: Feed the target as the next input decoder_input = target_tensor[:, i].unsqueeze(1) # Teacher forcing else: # Without teacher forcing: use its own predictions as the next input _, topi = decoder_output.topk(1) decoder_input = topi.squeeze(-1).detach() # detach from history as input decoder_outputs = torch.cat(decoder_outputs, dim=1) decoder_outputs = F.log_softmax(decoder_outputs, dim=-1) return decoder_outputs, decoder_hidden, None # We return `None` for consistency in the training loop why is it that "if target_tensor is not None": decoder_input = target_tensor[:, i].unsqueeze(1) but if target_tensor is None: _, topi = decoder_output.topk(1) decoder_input = topi.squeeze(-1).detach() specifically, isn't the shape of decoder_input different in both cases? I feel like in the first case, the shape of decoder_input is a 2D tensor but 1D in the second case thank you for your help Continue reading...
