These python programs were run on Spyder, to explore how BERT pretrained models are used with PyTorch.

BERT youtube comments.py uses the dataset available at https://archive.ics.uci.edu/ml/datasets/YouTube+Spam+Collection , and tries to classify comments as spam or ham.

BERT masked LM.py takes one sentence, sees its output in the pretrained corpus, and tries to finetune the corpus to give desired output.