TypeError callling tuner.best_estimator(best_training_job=best_job_name)

[jfelectron]

Describe the bug

Fetching best estimator from Tuner by name raises TypeError

To reproduce
A clear, step-by-step set of instructions to reproduce the bug.

1. Create XGBoost Estimator.
2. Create HyperparameterTuner with Extimator
3. After HPO finishes try to get best_estmator from tuner:
   best_job_name = self.tuner.best_training_job()
   best_estimator = self.tuner.best_estimator(best_training_job=best_job_name)

Expected behavior

Estimator is retrieved.

Screenshots or logs
2020-03-03T20:58:52.397Z best_estimator = self.tuner.best_estimator(best_training_job=best_job_name)
2020-03-03T20:58:52.397Z best_estimator = self.tuner.best_estimator(best_training_job=best_job_name)
File "/usr/local/lib/python3.7/site-packages/sagemaker/tuner.py", line 776, in best_estimator
2020-03-03T20:58:52.398Z training_job_name=best_training_job["TrainingJobName"],
2020-03-03T20:58:52.398Z TypeError: string indices must be integers

System information
A description of your system. Please provide:

• SageMaker Python SDK version: current
• Framework name (eg. PyTorch) or algorithm (eg. KMeans):
• Framework version:
• Python version:
• CPU or GPU:
• Custom Docker image (Y/N):

Additional context
Add any other context about the problem here.

[knakad]

Hi Jonathan!

Thanks for reaching out!
I updated the docstring on this method to be clearer. You need to pass in a python dict object containing a couple keys: #1331
I added an Example to the docstring as well to make it easier.

Hope that helps!

[knakad]

( https://sagemaker.readthedocs.io/en/stable/tuner.html will be updated on the next release, scheduled for Monday morning US Pacific time )