AzureML

The AzureML submodule contains utilities to connect to a workspace, train, tune and operationalize NLP systems at scale using AzureML.
For example, the DistributedCommunicator class defined in azureml_bert_util.py assists in making communication with multiple nodes for distributed training possible. azureml_utils.py contains a few helper functions that make it easy to authenticate, create, or retrieve an AzureML resource.