python

The CIFAR-10 and CIFAR-100 datasets are labeled subsets of the 80 million tiny images dataset. CIFAR-10 and CIFAR-100 were created by Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton. (Sadly, the 80 million tiny images dataset has been thrown into the memory hole by its authors. Spotting the doublethink which was used to justify its erasure is left as an exercise for the reader.)
The CIFAR-10 dataset
The CIFAR-10 dataset consists of 60000 32x32 colour images in 10 classes, with 6000 images per class. There are 50000 training images and 10000 test images.

The dataset is divided into five training batches and one test batch, each with 10000 images. The test batch contains exactly 1000 randomly-selected images from each class. The training batches contain the remaining images in random order, but some training batches may contain more images from one class than another. Between them, the training batches contain exactly 5000 images from each class.

Here are the classes in the dataset, as well as 10 random images from each:
airplane                                        
automobile                                        
bird                                        
cat                                        
deer                                        
dog                                        
frog                                        
horse                                        
ship                                        
truck                                        

The classes are completely mutually exclusive. There is no overlap between automobiles and trucks. "Automobile" includes sedans, SUVs, things of that sort. "Truck" includes only big trucks. Neither includes pickup trucks.
Download
If you're going to use this dataset, please cite the tech report at the bottom of this page.
Version    Size    md5sum
CIFAR-10 python version    163 MB    c58f30108f718f92721af3b95e74349a
CIFAR-10 Matlab version    175 MB    70270af85842c9e89bb428ec9976c926
CIFAR-10 binary version (suitable for C programs)    162 MB    c32a1d4ab5d03f1284b67883e8d87530
Baseline results
You can find some baseline replicable results on this dataset on the project page for cuda-convnet. These results were obtained with a convolutional neural network. Briefly, they are 18% test error without data augmentation and 11% with. Additi