This package contains the original AlexNet source code as it was in 2012, when it won the ImageNet competition. Geoffrey Hinton, Ilya Sutskever, and Alex Krizhevsky formed DNNResearch soon afterwards and sold the company, and the AlexNet source code along with it, to Google, which would continue work on it. This package also includes the parameter files trained on the ImageNet dataset.

Previously available as open source was Krizhevsky’s precursor to AlexNet, cuda-convnet, which was trained on the smaller CIFAR-10 dataset. While there are other existing repositories of code named "AlexNet" on the web, they are not the original code, but rather reimplementations based on the paper Krizhevsky, Sutskever, and Hinton published:

Krizhevsky, A., Sutskever, I. & Hinton, G. E. (2012). ImageNet Classification with Deep Convolutional Neural Networks. In F. Pereira, C. J. C. Burges, L. Bottou & K. Q. Weinberger (ed.), Advances in Neural Information Processing Systems 25 (pp. 1097--1105). Curran Associates, Inc.

In 2020, Alex Krizhevsky connected CHM to Geoff Hinton, who directed us to a team at Google. CHM worked with Google to identify the original 2012 version of the code and negotiate the terms of the public release as open source. CHM thanks Krizhevsky and Hinton for their support and David Bieber of Google DeepMind for his help in securing the release.

CHM is proud to present the source code to the 2012 version of Alex Krizhevsky, Ilya Sutskever, and Geoffery Hinton’s AlexNet, which transformed the field of artificial intelligence.