This repo includes a collection of notebooks that runs "Gan-Bert". There are some modifications to the code to adapt to notebook style execution and different datasets that were used
Original Gan-Bert Code - https://github.com/crux82/ganbert
Gan-Bert Paper: Croce, D., Castellucci, G., & Basili, R. (2020, July). Gan-bert: Generative adversarial learning for robust text classification with a bunch of labeled examples. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (pp. 2114-2119).
- Each individual notebook includes instructions on how to use them
- Gan_Bert_Colab_Clean.ipynb - is an attempt to run Gan-bert on notebook using the QC data provided in the original Gan-bert repo
- Gan_Bert_3_level_Sentiment_DataFrame_Colab.ipynb - a different dataset is used for three-level sentiment analysis. DataFrame format data was used as inputs. Data, stored as CSV, were loaded from a Google Drive path.