ImageNet (Deng et al. 2009) is an image database organized according to the WordNet (Miller 1995) hierarchy which, historically, has been used in computer vision benchmarks and research. However, it was not until AlexNet (Krizhevsky, Sutskever, and Hinton 2012) demonstrated the efficiency of deep learning using convolutional neural networks on GPUs that the computer-vision discipline turned to deep learning to achieve state-of-the-art models that revolutionized their field. Given the importance of ImageNet and AlexNet, this post introduces tools and techniques to consider when training ImageNet and other large-scale datasets with R. Now, in order to process ImageNet, we will first have to divide and conquer, partitioning the dataset into several manageable subsets. Afterwards, we will train ImageNet using AlexNet across multiple GPUs and compute instances. Preprocessing ImageNet and distributed training are the two topics that this post will present and discuss, starting with preprocessing ImageNet. When dealing with large datasets, even simple tasks like downloading or reading a dataset can be much harder than what you would expect.

Roundup Hello, here are a few bits of AI news for the weekend. You don't always need a ton of cash to buy a wad of GPUs to train your models super quickly. You can do it pretty cheaply on cloud platforms. There's also a robot that can play Where's Waldo (Wally in the UK), and Microsoft's computer that is trying to tell if you've found a joke funny. Public code for training ImageNet super quickly: A group of engineers have managed to train ImageNet to 93 per cent accuracy using hardware rented on public cloud platforms for just $40.

Note from Jeremy: I'll be teaching Deep Learning for Coders at the University of San Francisco starting in October; if you've got at least a year of coding experience, you can apply here. This is a new speed record for training Imagenet to this accuracy on publicly available infrastructure, and is 40% faster than Google's DAWNBench record on their proprietary TPU Pod cluster. Our approach uses the same number of processing units as Google's benchmark (128) and costs around $40 to run. The main training methods we used (details below) are: fast.ai's We used the classic ResNet-50 architecture, and SGD with momentum.