How do I get the name of Sagemaker Estimator's job

Question

I'm having a stumbling block with SageMaker. How do I know what my job name is?

For example:

mnist_estimator = MXNet(entry_point='/home/ec2-user/sample-notebooks/sagemaker-python-sdk/mxnet_mnist/mnist.py',
                        role=role,
                        output_path=model_artifacts_location,
                        code_location=custom_code_upload_location,
                        train_instance_count=1,
                        train_instance_type='ml.m4.xlarge',
                        base_job_name=’foo’,
                        hyperparameters={'learning_rate': 0.1})

and then when I call fit() it prints out:

INFO: Creating training job with name: foo-2018-01-10-20-13-57-893

and when I look in my S3 bucket I see:

2018-01-10 15:20:45 411784 artifacts/foo-2018-01-10-20-13-57-893/output/model.tar.gz

So my job name is "foo-2018-01-10-20-13-57-893" and not "foo", but how can I get this from python? I guess I'm looking for a way of extracting that from the Estimator itself, but I'm just blanking on it. I'm reading the python source but that's not helping me, probably because I'm just learning python also.

Answer 1

The _current_job_name contains the name of the job. So in the example from the question:

print(mnist_estimator._current_job_name)

would print

foo-2018-01-10-20-13-57-893