Tree Decomposition Attention for AMR-to-Text Generation

Implementation of the corresponding paper.

Credit: This code is based on the repo for the 2020 AAAI paper "Graph Transformer for Graph-to-Sequence Learning". We are grateful to the authors for open-sourcing their work.

Environment Setup

The code is tested with Python 3.6. All dependencies are listed in requirements.txt.

Data Preparation

The instructions to prepare AMR data are given in the generator_data folder.

Model Training and Evaluation

The following steps should be done in the generator folder. The default settings in this repo should reproduce the results in our paper. Please check all scripts for correct arguments before use.

1. Preprocess data and train

   sh prepare.sh  # vocab and data preprocessing
   sh train.sh
   

2. Test and postprocess

   sh work.sh  # test
   sh test.sh  # postprocess (make sure --output is set)
   

3. Evaluate

   ./multi-bleu.perl  # BLEU eval
   python chrF++.py -H [hyp] -R [ref]  # chrF++ eval
   java -Xmx2G -jar meteor-1.5.jar [hyp] [ref] -l en  # Meteor eval