Skip to content

Latest commit

 

History

History
25 lines (19 loc) · 761 Bytes

README.md

File metadata and controls

25 lines (19 loc) · 761 Bytes

Source Code Summarization with StructuralRelative Position Guided Transformer

Official implementation of our SANER 2022 oral paper "Source Code Summarization with StructuralRelative Position Guided Transformer".

You can download the full dataset from the link https://drive.google.com/file/d/1mAxDZZ3eTffezOl7Fkzi8ey9ac3ePkEZ/view?usp=sharing .

Quick Start

Training

cd main
python train.py --dataset_name python --model_name YOUR_MODEL_NAME

Logs are stored in

../modelx/YOUR_MODEL_NAME.txt

Testing

python test.py --dataset_name python --beam_size 5 --model_name YOUR_MODEL_NAME

Acknowledgement:

The implementation is based on https://github.com/wasiahmad/NeuralCodeSum and https://github.com/gingasan/sit3.