Current Build Status: Compiling, model does not work well with custom questions
Run python3 src/main.py, it will take around 5 minutes to load the models, then it will be ready to handle questions.
It will ask What is your question? and then spit out a prediction array for each likelihood array where each index correlates to an existing question in the database.
Example:
What is your question?
> Is piped text compatible with Web Services?
> [0.12, 0.00, ... 0.92, 0.18, 0.02]