th> c100 = torch.load('cifar100-train.t7')
th> print(c100)
{
        labelCoarse : ByteTensor - size: 50000
        data : ByteTensor - size: 50000x3x32x32
        label : ByteTensor - size: 50000
}

th> c10 = torch.load('cifar10-train.t7')
th> print(c10)
{
        data : ByteTensor - size: 50000x3x32x32
        label : ByteTensor - size: 50000
}