CIFAR10-Image-Classification-using-ResNet

Classify CIFAR10 images using Residual Network

Requirements

Python 3.5.1+ (not tested with Python2) Packages:

• keras 2.1.5+
• numpy
• matplotlib
• sklearn.metrics

Run Instructions

You can download the files or open it with your jupyter notebook/google colaboratory.

• Run Cifar10Classifer.ipynb

Results

Train accuracy - 90.86%           Train loss - 39.54%
Validation accuracy - 86.08%      Validation loss - 58.36%
Predictions on Test set - 92% success rate