README.md

Setup

Configure the Python virtualenv:

python3 -m venv env/

Activate the virtualenv and install all the dependencies:

source env/bin/activate;
pip install -r requirements.txt

Activating the Python virtual environment

whenever you go into the project folder, before you do anything you should activate the project's virtual environment by doing:

source env/bin/activate

To deactivate the virtual environment, simply type:

deactivate

Training

To train the model, simply run:

make train

Testing the model with random phrases

If you run the main.py file, you can type random phrases to see what the model outputs for them.

rasa shell nlu

Evaluating

You can write all the test phrases you want in the nlu_test.md file. After you have enough phrases to test, simply run:

make evaluate

The evaluation command will save the graphs in the hist.png, confmat.png and errors.json (if there's any errors) that you can consult or save if you want.

Cleaning temporary files

To remove all temporary created files and start from scratch, run:

make clean

It will remove the models folder and all the evaluation files such as hist.png, confmat.png, errors.json and the chatito generated datasets.