"ValueError: too many values to unpack (expected 2)" is occurred in windows local mode

[xnaiman]

System Information

• Framework (e.g. TensorFlow) / Algorithm (e.g. KMeans): Scikit-Learn
• Framework Version: 0.20.0 (official sagemaker-scikit-learn-container)
• Python Version: 3.6
• CPU or GPU: CPU
• Python SDK Version: 1.26.0
• Are you using a custom image: No

Describe the problem

When I execute fit method in local mode on windows, "ValueError: too many values to unpack (expected 2)" is occurred.

Cause

I already know that the cause is the difference between windows and linux drive description. Therefore, I specify the cause.

The following code is provided for sagemaker-python-sdk/src/sagemaker/local/image.py.

• class: _SageMakerContainer
• method: retrieve_artifacts

host_dir, container_dir = volume.split(':')

When this code is executed on windows, if volume is as follows.

C:\Users\oracle7\AppData\Local\Temp\tmp9sd97b87\algo-1-gnnm3\input:/opt/ml/input

The retun value is three and is as follows.

• C:
• \Users\oracle7\AppData\Local\Temp\tmp9sd97b87\algo-1-gnnm3\input
• /opt/ml/input

Thus, if platform.system () is Windows, three return values ​​should be assumed.

Minimal repro / logs

• Logs

algo-1-gnnm3_1 | 2019-06-13 00:45:24,804 sagemaker-containers INFO Reporting training SUCCESS
tmp9sd97b87_algo-1-gnnm3_1 exited with code 0
Aborting on container exit...
Traceback (most recent call last):
 File "c:\Users\oracle7\.vscode\extensions\ms-python.python-2019.5.18875\pythonFiles\ptvsd_launcher.py", line 43, in <module>
 main(ptvsdArgs)
 File "c:\Users\oracle7\.vscode\extensions\ms-python.python-2019.5.18875\pythonFiles\lib\python\ptvsd\__main__.py", line 434, in main
 run()
 File "c:\Users\oracle7\.vscode\extensions\ms-python.python-2019.5.18875\pythonFiles\lib\python\ptvsd\__main__.py", line 312, in run_file
 runpy.run_path(target, run_name='__main__')
 File "C:\Anaconda3\envs\sagemaker\lib\runpy.py", line 263, in run_path
 pkg_name=pkg_name, script_name=fname)
 File "C:\Anaconda3\envs\sagemaker\lib\runpy.py", line 96, in _run_module_code
 mod_name, mod_spec, pkg_name, script_name)
 File "C:\Anaconda3\envs\sagemaker\lib\runpy.py", line 85, in _run_code
 exec(code, run_globals)
 File "c:\Users\oracle7\Documents\GitLab\System\predictive-maintenance\sagemaker\tutorial_basic\train.py", line 24, in <module>
 sklearn.fit({'train': train_input})
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\estimator.py", line 234, in fit
 self.latest_training_job = _TrainingJob.start_new(self, inputs)
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\estimator.py", line 592, in start_new
 estimator.sagemaker_session.train(**train_args)
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\session.py", line 317, in train
 self.sagemaker_client.create_training_job(**train_request)
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\local\local_session.py", line 73, in create_training_job
 training_job.start(InputDataConfig, OutputDataConfig, hyperparameters, TrainingJobName)
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\local\entities.py", line 69, in start
 self.model_artifacts = self.container.train(input_data_config, output_data_config, hyperparameters, job_name)
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\local\image.py", line 143, in train
 artifacts = self.retrieve_artifacts(compose_data, output_data_config, job_name)
 File "C:\Anaconda3\envs\sagemaker\lib\site-packages\sagemaker\local\image.py", line 239, in retrieve_artifacts
 host_dir, container_dir = volume.split(':')
ValueError: too many values to unpack (expected 2)

• Exact command to reproduce:

sklearn = SKLearn(
 entry_point='scikit_learn_iris.py',
 train_instance_type="ml.c4.xlarge",
 role=role,
 sagemaker_session=sagemaker_session,
 hyperparameters={'max_leaf_nodes': 30})

sklearn.fit({'train': train_input})

[mvsusp]

Hi @xnaiman ,

We currently don't support Local mode for windows instances. This feature is already included in our roadmap.

Sorry for any inconvenience about that.

[xnaiman]

@mvsusp

This feature is already included in our roadmap.

Where can I see the roadmap(plan) ?

[mvsusp]

Unfortunately, the roadmap is not public available yet, @xnaiman .

[manvento]

Hi,
I have proposed a possibile solution for this issue. For me it works good.
You can find it here:
#1125

[linehammer]

ValueError is a standard Exception raised by various methods that perform range-checking of some kind to signal that a value provided to the method fell outside the valid range. Python functions can return multiple variables . These variables can be stored in variables directly. This is a unique property of Python , other programming languages such as C++ or Java do not support this by default. The valueerror: too many values to unpack occurs during a multiple-assignment where you either don't have enough objects to assign to the variables or you have more objects to assign than variables.

[nadiaya]

#1125 has been merged. Resolving the issue.