About the sequence length and sentence length.

declare-lab / MELD

MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversation

GNU General Public License v3.0

788 stars 200 forks source link

About the sequence length and sentence length. #6

Closed KuzmaNg closed 5 years ago

KuzmaNg commented 5 years ago

Hi, Meld crew. I tried running baseline code with text-only sentiment classification, and it worked. However, I got one question in baseline.py (Line 124). It's about the input_length in Embedding layer. I think it should be the sentence_length instead of the sequence_length, since the 2nd dimension of the _concatenatedtensor is a negative number (-48) in my case.

devamanyu commented 5 years ago

Hi. Thank you for noticing the error. We have now corrected the line!

KuzmaNg commented 5 years ago

@devamanyu Thank you.