idea from:
- https://medium.com/analytics-vidhya/creating-a-very-simple-u-net-model-with-pytorch-for-semantic-segmentation-of-satellite-images-223aa216e705
- https://medium.com/analytics-vidhya/how-to-create-a-custom-dataset-loader-in-pytorch-from-scratch-for-multi-band-satellite-images-c5924e908edf
Kaggle directories are:
- /kaggle/input (This is there the dataset is located)
- /kaggle/working (his is the workspace directory)
- /kaggle/weights (this is where we save the weights .pt)