1. Fine tuning a pre-trained BERT transformer model with hand labeld text segements for sentiment analysis.
2. Developement of a visual debugging tool for this model or more generel:

A visual debugging tool to fine-tuned BERT models for (multilabel) sequence classification tasks

Fine-tuned BERT model and UMAP model have to be downloaded separately here. Place the .pk files in the data/ folder, the model in the model folder or configure the model folder respectively in the config.py file.