This project explores the effectiveness of the LeNet model on the CIFAR10 dataset. It serves as a base line for future experimentation.
To run the LeNet model written in tensorflow, change directories into the tensorflow folder and run leNet.py using python. The LeNet model will be trained on 20 epochs of the training data with batch sizes of 100. Every 100 batches the cost and test accuracy will be printed. After one epoch has completed the test accuracy will be printed.
To run the LeNet model written in Keras, change directories into the keras folder and run the run_lenet.py file. Epoch and batch size are 20 and 100 respectivly.