This repository is based on the code. Different from their work, this repository focuses on the DRG-to-text generation task. Another repository is its sister, which includes the poster for CLIN 32 conference.

Environments and Dependencies

• Python 3.6
• PyTorch 1.5.0

Datasets

In our experiments, we use the following datasets same as sister repositoty :

Preprocess

First, convert the dataset into the format required for the model.

sh preprocess_nl.sh   

Training

For training the model using the DRG dataset, execute:

./train_sh.sh <gpu_id> <gnn_type> <gnn_layers> <start_decay_steps> <decay_steps>

Options for <gnn_type> are ggnn, gat or gin. <gnn_layers> is the number of graph layers. Refer to OpenNMT-py for <start_decay_steps> and <decay_steps>.

We lower the learning rate during training, after some epochs, as in Konstas et al. (2017).

Examples:

sh train_nl.sh 0 gin 2 5000 5000

Decoding

For decode on the test set, run:

sh decode_nl.sh <gpu_id> <model> <nodes_file> <node1_file> <node2_file> <output>

Example:

sh decode_nl.sh 0 model_ggnn.pt test.txt-src-nodes.txt test.txt-src-node1.txt test.txt-src-node2.txt output-ggnn-test.txt

Cite

@article{wang-2022-dutch-compare,
  title={Comparing Neural Meaning-to-Text Approaches for Dutch},
  author={Wang, Chunliu and
      Bos, Johan},
  journal={Computational Linguistics in the Netherlands Journal},
  volume={12},
  pages={269–286},
  year={2022}
}