/cnn-cifar10-example

Simple example of a Convolutional Neural Network trained with Tensorflow on the CIFAR-10 dataset

Primary LanguageJupyter Notebook

CNN image classification example

This is a simple example of image classification using a CNN and the CIFAR-10 dataset.

The example is implemented using Tensorflow with relatively low level functions to better understand what happens in detail.

This example is taken to a good amount from github.com/GzuPark/image_classification with some minor code changes and some model architecture changes to improve the total accuracy. The final accuracy is roughly 80%, compared to about 65% for the initial model.

It must be noted that the model could of course be tweaked to get well over 90% of accuracy. See here for the best models.