The CIFAR-10 dataset is a collection of 60,000 32x32 color images in ten
different categories.  The dataset was created specifically to use for training
and testing image classification algorithms.  The categories are:

  * Airplane    * Automobile
  * Bird        * Cat
  * Deer        * Dog
  * Horse       * Frog
  * Ship        * Truck

For TripleBlind examples, we have split this collection into two datasets of
25,000 images for collaborative training, plus one more dataset of 10,000
images for testing.  Each dataset is placed on an independent Access Point to
represent independent organizations with similar data who wish to jointly build
a network without sharing their data.

Ultimately this outputs the datasets:
 * "EXAMPLE - CIFAR-10, part 1"
 * "EXAMPLE - CIFAR-10, part 2"

And an equivalent but much smaller group:
 * "TEST - CIFAR-10, part 1"
 * "TEST - CIFAR-10, part 2"

There is one additional testing files that gets created at this time.
This is stashed in Google Drive for download when needed:
 * test_cifar.zip (10 images)
