Deep Learning in the Cloud with MATLAB R2016b

MATLAB code for Deep Learning in the Cloud with MATLAB R2016b white paper.

This provides support code for the " Deep Learning in the Cloud with MATLAB R2016b" white paper:
https://www.mathworks.com/content/dam/mathworks/mathworks-dot-com/products/files/super-computing-deep-learning-white-paper.pdf

This presents a worked example of how to train a neural network to perform image classification on the CIFAR-10 data set. Using the Parallel Computing Toolbox and MATLAB Distributed Computing Server for Amazon EC2, multiple networks are trained in parallel using GPU clusters in the cloud. The relative accuracy and memory footprints of the different networks is compared to chose the best network architecture for the problem.

To replicate the results of the paper run the cifar10DeepLearning.m script. This will train 16 different neural networks in parallel. Refer to the white paper for help with:
- Creating a cloud cluster.
- Uploading and preparing the CIFAR-10 data set on your cluster.

The following helper functions are also provided to help users upload, manage, and backup data on their cluster:
- copyToCluster.m
- deleteFromCluster.m
- existsOnCluster.m
- createSnapshot.m
For more information about using these functions refer to their help text.