Microsoft Azure offers a service called Batch AI that allows us to run our machine learning models on clusters of Azure virtual machines.
Running Distributed TensorFlow on Microsoft Azure
Getting ready
As the first step we need to have an Azure account: if you do not already have one, you can create one for free here: https://azure.microsoft.com/en-us/services/batch-ai/. The Azure offers $200 credit for 30 days to new users. This recipe will follow the example provided by Microsoft Azure to run MNIST on two GPUs using TensorFlow distributed, the relevant codes are available on Github: https://github.com/Azure/batch-shipyard/tree/master/recipes/TensorFlow-Distributed