Some simple scripts to train a Keras model using ResNet50.
Use DirectoryTrainer.py to train a model with folders of training images and validation images.
Use DirectoryTester.py to test a model with a folder of unseen images.
Use Logloss.py to calculate the log loss between output from DirectoryTester.py and a known solution.
Use Ensemble.py to create an ensemble from several outputs of DirectoryTester.py.
Used for Kaggle competitions successfully.