Intent classifcation using BERT on 20 classes of CLINC150 dataset
- Dataset
- Folder with fine tuned weights and tokenizer (output_dir folder).
- Jupyter notebook with the code. Run on Google Colaboratory.
- Class encapsulating the preprocessing, training, testing, evaluating pipeline is intent_classification.py
- requirements.txt auto-generated using the terminal. Some packages may have been pre-installed on Google Colaboratory and not need for this project.