This repository is dedicated to fine-tune the text classification models. It primarily focuses on fine-tuning the pre-trained BERT model, utilizing the ratsnlp package.
If you want to try it in the colab, please refer to notebook here.
You will need to prepare a dataset comprising two columns: one for the document and the other for label. An example of the dataset format is as follows:
| label | document |
|---|---|
| sadness | I'm so sad |
| happiness | I'm happy!! |
For a more detailed understanding, please refer to the example dataset.
This repository includes a very small sample example dataset sourced from Kaggle, available here: Kaggle Dataset
Note: The sample in the repo is very small size, it is recommended to prepare a much larger dataset.