CIFAR-10

CIFAR-10 dataset

The CIFAR-10 dataset

Dataset Statistics

  1. Color: RGB
  2. Sample Size: 32x32

The number of categories of CIFAR-10 is 10, that is airplane, automobile, bird, cat, deer, dog, frog, horse, ship, truck.

The Number of Samples per Category for CIFAR-10

Category airplane automobile bird cat deer dog frog horse ship truck Total
#Training Samples 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 50,000
#Testing Samples 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 10,000

Samples

CIFAR-10 Sample

Dataset Usage

TensorFlow:

TensorFlow @TensorFlow_Convolutional_Neural_Networks

Caffe:

export CAFFE_ROOT='path_to_caffe_root_folder'
cd $CAFFE_ROOT
./data/cifar10/get_cifar10.sh
./examples/cifar10/create_cifar10.sh

Caffe @caffe_Alex’s_CIFAR-10_tutorial_Caffe_style

Dowload CIFAR-10 for Caffe

Torch:

th train-on-cifar-10.lua

General tools for Python

Official Python function:

def unpickle(file):
    import cPickle
    with open(file, 'rb') as fo:
        dict = cPickle.load(fo)
return dict