We make use of a set of helper functions during the build-up to our training script. In this section, we'll briefly cover these functions and the steps needed to make simGAN train:

1. Create a method for load_h5py to load the simulated and real eye gaze data:

     def load_h5py(self,data_path):
        with h5py.File(data_path,'r') as t_file:
            print('Images found:',len(t_file['image']))
            image_stack = np.stack([np.expand_dims(a,-1) for a in t_file['image'].values()],0)
        return image_stack