CNN models. To run the python files. put balancedData_shuffled, *.py, and *_weights.h5 under the same directory. The result of training, validation and testing are images names "accuracy.png" and "loss.png"

1. Compile dense121_train.py.py read images at '/projectnb/cs542sp/idc_classification/data/' process the Dataset and shuffle it, name as "balancedData_shuffled" then store it at "./balancedData_shuffled" Train the model and get the weights called "densenet_weights.h5".
2. Compile dense121_pred.py load dataset called "./balancedData_shuffled" load weights for the model ('densenet_weights.h5')

return the predicted result.

The file newvgg.py is our final model.

1. read dataset at "./balancedData_shuffled"
2. output weights of the model: 'newvgg_weights.h5'

python alex_net.py