View Code? Open in Web Editor
NEW
EMNLP2021-Syntactically-Informed Unsupervised Paraphrasing with Non-Parallel Data
sup's Introduction
๐ Hi, Iโm @lanse-sir
๐ Iโm interested in natural language processing and controllded text generation.
๐ฑ Iโm currently learning ...
๐๏ธ Iโm looking to collaborate on ...
๐ซ How to reach me ...
sup's People
Contributors
Watchers
sup's Issues
xxx_parse.txt and xxx_template.txt is required for train_cvae.py, so how to process a new dataset to get them.
For instance:
code for parsing sentences into syntactic trees and keep the top 3 layers
code for building the vocab file
code for training the syntactic seq2seq neural parser evaluator
file para1m_fine-tune.yaml
and so on......
I should be most grateful if the missing codes and files needed to run the model can be released.
The model.py file under autocg executes from autocg.decoder.rnn_decoder import RNNDecoder
๏ผ but there is not decoder directory
Hi,
Could you please provide the dataset that you use in your job? or the links to download?
Thank you very much!
Since we can't find the fine-tune configuration file, we modify the configuration file of train_cave slightly for fine-tune, but the self loss keeps rising, can you provide the fine-tune configuration file