Dependencies that can be installed through PyPI are

• Tensorflow
• Keras
• NumPy
• SciPy
• scikit-image
• MatPlotLib
• tqdm

In addition, the following packages are also required

• Pytorch
• PyJet

Place the testing data into a folder on the top level of the repository called input/test and the training data in input/train in the same format as Kaggle's Data Science Bowl competition's datasets.

Run the command from the src directory

python3 run_model.py train-unet-1 --test

to generate a submission. The submission can be found in the submissions folder under the name train-unet-1.csv