spam_classifier.ipynb
Bi-directional LSTM spam classifier model trained on dataset of about 5000 mail. The model uses trained embeddings generated using fasttext library with skipgram method. The notebook a step by step illustration to how fasttext is used to train embeddings and how to use these embbedings in a keras embedding layer.
spam classifier.py
a naive model that uses traditional text preprocessing and basic machine learning algorithms to identify spam mails