A simple convolutional neural network for CIFAR10 dataset in tensorflow
Files:
main.py - Includes the final training and testing code
model.py - Includes just the model skeleton (i.e. the neural network)
train.py - Includes the loss function, training and evaluation code
helper_data.py - Includes data preprocessing code and the data batch class
helper_nn.py - Includes code required to instantiate the model
Dependencies:
tensorflow - used r0.11 to write and test the code
numpy: basic matrix operations
scipy: for creating image distortions
argparse: for command-line inputs
h5py: for reading in the CIFAR10 data in hdf5 format