Overview of things learned:
- Technical Area:
- Learned about advanced models like BERT, xlm, roberta, distilbert
- Learned how to use SimpleTransformers library in the model
- Gained an understanding of combining an advanced model (BERT) and a simple one (logistic regression)
- Used matplotlib to plot some graphs about the data and learned about balancing data
- Tools:
- Jupyter Notebooks
- Google Colab
- SimpleTransformers
- Pandas, numpy, csv, matplotlib
- Soft Skills:
- Enhanced my knowledge about advanced machine learning
- Collaborated with my teammates for presenting our work to our mentor
- Extensively used various online resources to debug my code
Achievements:
- Successfully created a recommender system that uses BERT and logistic regression to predict the category of a post (topic title and leading comment was input)
- Gained a decent understanding of advanced models like BERT, xlnet, roberta
- Understood what steps to take next for solving the problem of unbalanced data
Challenges:
- Running the model took a lot of time, and I had to debug it repeatedly
- The model takes up a lot of RAM, so it crashed a few times