/data2text-seq-plan-py

Code for TACL 2022 paper on Data-to-text Generation with Variational Sequential Planning

Primary LanguagePythonMIT LicenseMIT

data2text-seq-plan-py PWC

This repo contains code for Data-to-text Generation with Variational Sequential Planning (Ratish Puduppully and Yao Fu and Mirella Lapata; In Transactions of the Association for Computational Linguistics (TACL)); this code is based on an earlier (version 0.9.2) fork of OpenNMT-py.

Citation

@article{puduppully-2021-seq-plan,
  author    = {Ratish Puduppully and Yao Fu and Mirella Lapata},
  title     = {Data-to-text Generation with Variational Sequential Planning},
  journal = {Transactions of the Association for Computational Linguistics (to appear)},
  url       = {https://arxiv.org/abs/2202.13756},
  year      = {2022}
}

Requirements

All dependencies can be installed via:

pip install -r requirements.txt

Code Details

The steps for training and inference for the MLB dataset are given in README_MLB.

Model

The links for the models are MLB, RotoWire and German RotoWire.

Model Outputs

The model outputs are at MLB and German-RotoWire.

Acknowledgements

Part of the code is based on the Sequential Knowledge Transformer repo.