Daniel Aloni (ID. 207983875), Nitzan Nehushtan (ID. 315870956) Submitted as final project for the NLP course, IDC, 2022
This project introduces a discriminative language model aimed to predict functions names given functions bodies and signatures.
Running the code is as simple as running main in the python_model.py file. Data is already preprocessed and stored in the data/ folder.
- Collect Python stdlibs.
- Collect JS stdlibs.
- Implement Super preprocessor class.
- Go over design.
- Implement Py/JS preprocessors.
- Implement Code2Vec.
- Build model by chosen architecture.
-
- Train & evaluate.
-
- Post-processing.