This is a NLP deep learning Python program using Keras / Tensorflow 2.
It uses a bi-directional transformer model for natural language processing.
Coded in a professional style to meet PEP8 standards. You can gain a greater understanding of deep learning transformer NLP structure by reviewing the code and comments.
Currently structured using a JSON input file to match sentences with the appropriate intent. Hence the intent determination logic coded within. More file upload options will be added later.
Basic JSON format and a few examples are given. Actual JSON file is much larger for production use.
Uses the following Python modules: re - regular expressions Numpy Tensorflow 2 Keras matplotlib JSON poetry
easy install w Pycharm and Poetry Open your project create a poetry environment in PyCharm
then in your terminal console in Pycharm: $ python -m poetry lock --no-update $ python -m poetry install
key packages: $ poetry add numpy $ poetry add matplotlib $ poetry add tensorflow $ poetry add tensorflow_datasets
check tensorflow was added correctly by: $ python -m pip show tensorflow