Config.py - The hyperparameters of the network customer_init.py - some initialization method Model.py - abstract model readfile.py - read a batch of images from file and send it to tensorflow to train (refer to Tensorflow Dataset) SegnetModel - implementation of segnet model util.py - util methods

1. Make several dirs under the data folder: -- data

   -- Logs -- train -- val -- submission

2. download data from kaggle website

3. put the training data under train folder and the test data under val folder

4. You can train your model by writing code in the main function of SegnetModel

    if __name__ == "__main__":
        segmodel = SegnetModel()
        segmodel.training()  

EDA: https://colab.research.google.com/drive/12HpbXiewjQdJaZ3qm1QFEBS3xm_lvYbh#scrollTo=cmWKLEOlQSVo training: https://github.com/JasonChu1313/Satellite-Segmentation/tree/master/src