Azure Cloud

Azure Cloud is another environment for training models in addition to HPC, but it offers much more than just training. Azure Cloud provides a development environment that supports:

  • Dataset versioning: Data could be imported and versioned
  • Model versioning: Models could be saved and versioned
  • Code versioning: Code changes could be saved and versioned

https://docs.microsoft.com/en-us/azure/architecture/reference-architectures/ai/training-python-models

Using Azure ML service, a model could be trained on a compute, or a training cluster if necessary. A model could be deployed to an inference cluster for inference, or deployed to another system to be comsumed, such as WebFocus for dashboards, charts, and reports.