A Convolutional Neural Network to classify CIFAR-10 image dataset using TensorFlow.

To open the main code, simply open tf-cifar10.ipynb on any desktop browser, or you can download and run the cells in a Python 3 environment. The code is presented in a Jupyter Notebook / iPython notebook for readability purposes.

This project classifies images from the CIFAR-10 dataset. The dataset consists of airplanes, dogs, cats, and other objects. We will preprocess the images, then train a convolutional neural network on all the samples. The images need to be normalized and the labels need to be one-hot encoded.

The dataset is broken into batches to prevent our machine from running out of memory. The CIFAR-10 dataset consists of 5 batches, named data_batch_1, data_batch_2, etc.. Each batch contains the labels and images that are one of the following: airplane automobile bird cat deer dog frog horse ship truck

This code builds combination of convolutional neural network, max pooling, dropout, and fully connected layers. At the end, there will be a test on the neural network's predictions on sample images.

The final accuracy is about 67%, which is much better than pure guessing (pure guessing on CIFAR-10 dataset will stochastically yield 10% accuracy).

This project requires Python 3 and the following Python libraries installed:

• TensorFlow
• NumPy
• tqdm - Progress Meter
• pickle
• helper