To install the repository, ensure that you enable your virtual environment. I can recommend py-venv. This can easily be installed on Unix-like OS (Mac, Linux). For Mac, use Brew, and for Linux use the suitable package manager - i.e. pacman for Arch-linux, apt-get for Ubuntu, etc.
Then create your virtual environment using:
python -m venv geoAI
Then activate your virtual environment using:
source geoAI/bin/activate
Thereafter, simply run:
pip install -r requirements.txt
to install all of the necessary dependencies.
To run any of the CNN models, please call the following command:
python src/<the_cnn_you_want_to_run.py>
where <the_cnn_you_want_to_run>={cnn.py, cnn_data_augment.py, cnn_binary.py}.
It should generate a directory called results/, wherein runs will be saved according to your local time.
An example output of src/cnn_data_augment.py:
Which yields an MSE loss