shashwath94 / hierarchical-seq2seq Goto Github PK

A PyTorch implementation of the hierarchical encoder-decoder architecture (HRED) introduced in Sordoni et al (2015). It is a hierarchical encoder-decoder architecture for modeling conversation triples in the MovieTriples dataset. This version of the model is built for the MovieTriples dataset.

Python 100.00%

nlp deep-learning pytorch seq2seq-pytorch hred

hierarchical-seq2seq's People

Contributors

Stargazers

Watchers

Forkers

leidongfeng ammieqi llluoling seanggggggg

hierarchical-seq2seq's Issues

word_summary pickle file

Can you please specify what does word_summary file contains? Also, can you please provide the dataset. I have already mailed the concerned author but has got no reply from them.

KeyError: tensor(8) in util.py for making sentence for idx

I am trying out this code and could successfully train the code but when testing it I get this error of KeyError: tensor(8) in utils.py for the line sent.append(inv_dict[i]). Could anyone help.

what mean the "Language model objective"

Excuse me,
Thanks for your work on implement this code, and after reading your code, i'm confused about the two commands meansing as follows:

For Bi-HRED + Language model objective (inverse sigmoid teacher forcing rate decay)
For Bi-HRED + Language model objective (Ful teacher forcing)

can you explain what the meaning of lm, why there are another one gru in Decoder.py if the options.lm is true?Looking forward your patient explaination.

Recommend Projects