Whats the difference of using a pad sequence from torch.nn.utils.rnn.pad_sequence()
or a PackedSequence
in the torch.nn.LSTM()
Only computation? or also relates to the result.
Whats the difference of using a pad sequence from torch.nn.utils.rnn.pad_sequence()
or a PackedSequence
in the torch.nn.LSTM()
Only computation? or also relates to the result.