Error when running hyperparameter tuning job locally

[Srijstha]

Describe the bug Running a hyperparameter tuning job locally using a sample code as given below produces an error: AttributeError: 'LocalSagemakerClient' object has no attribute 'create_hyper_parameter_tuning_job'.

tuner = HyperparameterTuner(estimator,
objective_metric_name,
hyperparameter_ranges,
metric_definitions,
max_jobs=4,
max_parallel_jobs=2,
objective_type=objective_type,
base_tuning_job_name="hpo-tuning-demo"
)

tuner.fit(inputs=channels)

To reproduce Use the Amazon AWS example from the link: https://github.com/aws/amazon-sagemaker-examples/tree/master/hyperparameter_tuning/keras_bring_your_own, and change instance_type to 'local'. For Hyperparameter tuning job, instead of using configurations as in the example, just use HyperparmeterTuner class as shown above.

Expected behavior I'd expect it to work without the error.

Screenshots or logs Screenshot of the error output given below.

System information A description of your system. Please provide:

• SageMaker Python SDK version: 2.15
• Framework name (eg. PyTorch) or algorithm (eg. KMeans): TensorFlow
• Framework version: 2.3.1
• Python version: 3.8
• CPU or GPU: CPU
• Custom Docker image (Y/N): Y (Dockerfile from the AWS example code)

Additional context Instead of using the AWS instances, trying to use the SageMaker hyperparameter tuning job locally for testing purposes.

[chuyang-deng]

Hi @Srijstha, SageMaker Python SDK local mode currently does not provide tuning job functionalities. Sorry for the inconvenience.

[Srijstha]

Hi ChuyangDeng, Thank you for your quick response. I hope that the feature will be added soon so that people will benefit from using it locally.

[idanmoradarthas]

@chuyang-deng Any news? Even running HyperparameterTuner with only one job in local would be a great help.

[espoirMur]

I have the same issue here...

[akshika47]

Hi guys,

Any update on this?