/MNIST-CIFAR-10

Convolutional Neural network for MNIST digits, CIFAR-10, and a clothing database.

Primary LanguageJupyter Notebook

CNN for MNIST/CIFAR-10

Using Jupyter Notebook, I created a two layer convolutional neural network for MNIST digits (recognizing handwritten digits from 0-9) and for CIFAR-10, a database of images for 10 generic objects (airplane, ship, cat, etc).

I then applied this for a new data-base for fashion related objects such as shoes and t-shirts.

Over multiple epochs of training, I eventually achieved an accuracy of 97.4%.