lfy79001 / reghnt Goto Github PK

Code for COLING 2022 long paper: Answering Numerical Reasoning Questions in Table-Text Hybrid Contents with Graph-based Encoder and Tree-based Decoder

Python 100.00%

reghnt's People

Contributors

Stargazers

Watchers

Forkers

ohsuz castillosebastian

reghnt's Issues

there is no cache/reghnt_roberta_large_cached_train

FileNotFoundError: [Errno 2] No such file or directory: 'reg_hnt/cache/reghnt_roberta_large_cached_train.pkl'

the code about the span extraction tree

In your paper,( Appendix C Details of the span extraction question), but I can't find the code related to the span extraction tree, can you specify which file it is in? Thanks!

I can't reproduce the results of the experiment

I used the given training script to train the model on a single V100. The experimental results I obtained on the dev set (EM/F1=72.1/79.2 & Scale=93.8) lagged behind those reported in the paper (EM/F1=73.6/81.3 & Scale=95.3). I noticed that the code of span extraction tree hasn't been uploaded but you say it is messy and has little effect on performance. So is there any detail that prevents me from reproducing the results of the experiment?

Recommend Projects