The gist of this project is to provide a rough stratregy to detect spam text content utilizing LSTM neural networks. Structure of the said neural network is as follows:

The goal of this implementation is to provide sufficient accuracy of spam text detection (~0.98 accuracy on the test portion of the data)

A simple solution (akin to Hashcash) is provided to limit spam email traffic - by making use of PoW before a client (potential spammer) sends an email.

$ python spam_filter.py [path_to_weights_file]

If no weights are supplied, the application enters training mode

Run the tests with pytest