This repository stores binary data that is consumed by tests in the dotnet/machinelearning repository.
MICROSOFT PROVIDES THE DATASETS ON AN "AS IS" BASIS. MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, GUARANTEES OR CONDITIONS WITH RESPECT TO YOUR USE OF THE DATASETS. TO THE EXTENT PERMITTED UNDER YOUR LOCAL LAW, MICROSOFT DISCLAIMS ALL LIABILITY FOR ANY DAMAGES OR LOSSES, INLCUDING DIRECT, CONSEQUENTIAL, SPECIAL, INDIRECT, INCIDENTAL OR PUNITIVE, RESULTING FROM YOUR USE OF THE DATASETS.
The datasets are provided under the original terms that Microsoft received such datasets. See below for more information about each dataset.
MNIST data originally from NIST and modified by Chris Burges, Corinna Cortes, and Yann LeCun. http://yann.lecun.com/exdb/mnist/
More information: https://en.wikipedia.org/wiki/MNIST_database
More information: https://www.cs.toronto.edu/~kriz/cifar.html
Originally released in 2016, developed by researchers at DeepScale, University of California, Berkeley, and Stanford University.
More information: https://en.wikipedia.org/wiki/SqueezeNet, https://github.com/onnx/models/tree/master/squeezenet