Until now, we saw how to set up MXNet in a local system. MXNet is also supported by various cloud platforms such as AWS, Microsoft Azure, and GCP.

For working with MXNet using AWS, we can use Amazon SageMaker, which provides a fully-fledged platform to build, train, and deploy deep learning models in a scalable manner. You can also leverage the AWS deep learning AMIs like the NVIDIA deep learning AMIs, which are pre-configured environments to prototype deep learning applications quickly. In AWS, it is also possible to build our own customized deep learning environments for using MXNet.

GCP provides the NVIDIA GPU Cloud Image that offers an optimized environment to run GPU optimized containers for running deep learning applications using MXNet.

Microsoft Azure offers the NVIDIA NGC image for Deep Learning and HPC, which is an optimized environment for running the GPU-accelerated containers from the NGC container registry, which includes frameworks such as MXNet, CNTK, and Theano.