Sent the packed sequence into LSTM

[xdwang0726]

SU4MLC/semantic-unit-based/models/rnn.py

Line 38 in 477bcfe

outputs, state = self.rnn(embeds)

Hi, I am just wondering should it like send the packed_padded embedding to the LSTM first, and then pad the output which is getting from LSTM? (like in the referenced line, should Line38 be placed before Line37)
Thanks!