Batch AI Service (Retired)
Autor: Microsoft
Azure Batch AI provides job scheduling and resource management for AI applications.
Note: Azure Batch AI has been retired. You can no longer create new resources of this service. The capabilities of Azure Batch AI are now available as a managed compute target in Azure Machine Learning service. For more information, see What happened to Azure Batch AI?
Azure Batch AI is designed to help you run large AI training and testing workloads in the cloud. Batch AI supports multiple training toolkits such as Tensorflow, CNTK, Chainer, among others, as well as the ability to deploy your own software stacks at scale.
Simply describe your job requirements and configuration to run, and Batch AI will handle the rest.
With Batch AI you will be able to leverage the full power of Azure hardware including powerful GPU's such has the NVIDIA P40, P100, K80 and M60.
Highlights:
- Deploy the software you need using Docker containers or virtual machine images
- Scale up and down compute pools to support your work and manage costs
- Create an NFS share for training data and logs
- Mount Azure Files, NFS volumes, or parallel file systems to the virtual machines or containers