This is an algorithm designed to recognize the ten classes of objects in the CIFAR10 dataset. It is written in TensorFlow and inspired by the AlexNet architecture, which is cited below. An accuracy of about 77% was achieved, which is lower than the original AlexNet achieved. The authors of the AlexNet journal had access to several GPUs, whereas I am using a personal computer, which explains the discrepancy.
This project demonstrates my ability to read scientific literature in deep learning at the graduate level, as well as my ability to code the algorithms contained within this literature. Apart from the journal mentioned below, the code written in this Notebook is my own.
Krizhevsky, Alex, et al. “ImageNet Classification with Deep Convolutional Neural Networks.” Communications of the ACM, vol. 60, no. 6, 2017, pp. 84–90., doi:10.1145/3065386.