Machinelearningnotebooks: ImportError: No module named 'azureml.core'

[kritdikoon]

I am doing AML tutorial in 03.auto-train-models.ipynb.
As a very new user, I get stuck from the beginning.

From Prerequisites
I am creating a workspace and its configuration file (config.json)
I create python code and run in Azure ML workbench as below: (xxx is hidden info)

#

from azureml.core import Workspace
subscription_id ='xxxxxxx-xxxx-xxxx-bf6a-e03e4fe8bced'
resource_group ='xxxxxxx-AZURE-DEV'
workspace_name = 'xxxxxx-test-AML'

try:
ws = Workspace(subscription_id = subscription_id, resource_group = resource_group, workspace_name = workspace_name)
ws.write_config()
print('Library configuration succeeded')
except:
print('Workspace not found')

#

After I run the script, there is an error below:

File "config.py", line 1, in
from azureml.core import Workspace
ImportError: No module named 'azureml.core'

Could you help me what did I do wrong?
It seem that azureml library is not connected.
by the way, I have installed azureml and confirm with pip freeze. (azureml==0.2.7)

conda version : 4.5.11
python version : 3.6.5.final.0
platform : win-64

[wchill]

Hello, do you have the azureml-train-automl package installed? It's needed for 03.auto-train-models.ipynb and will also pull in the azureml.core package as a dependency.

[kritdikoon]

Hi, wchill

Many thanks for your answer.
I have tried installing azureml-train-automl package.
I have checked my package with pip freeze. This package show up [azureml-core==0.1.65] .
Is it the one I need?

By the way, I have run above python script again.
This error is still shown up.

#

File "config.py", line 1, in
from azureml.core import Workspace
ImportError: No module named 'azureml.core'

#

PS. I feel I am get closer to solution.
Do you think it is due to python version or conflicting library?
It seems there are some conflicting with Tensorflow package.

[wchill]

Yes, that is the correct package.

Just to verify, can you run pip install --upgrade azureml-sdk[notebooks,automl]? This should install all SDK packages and if it doesn't then we do have a problem.

Your Python version is fine. We do have Tensorflow locked down to version 1.10.0 for azureml-train-automl; maybe that's what's causing the conflict? Do you have any other packages installed in that environment that are taking a dependency on Tensorflow?

[wchill]

One more thing, it seems like you are using AzureML Workbench, can you try using the conda environment directly (in a command prompt)? This will help us identify where the issue is if there is one.

[rastala]

The new AzureML Python SDK and old AzureML Workbench are incompatible. To use the azureml-sdk, the recommended approach is to create a conda environment, or python virtual environment to ensure a clean installation with no dependency conflicts, or packages shadowing each other.

See here for documentation: https://docs.microsoft.com/en-us/azure/machine-learning/service/how-to-configure-environment

[kritdikoon]

Thank you wchill and rastala for very helpful answers.
I have set up a virtual environment to avoid conflicts.
I have passed the step and confirm my workspace account.

SOLUTION: set up virtual python environment

Yet, I got stuck with another check point which I may put in separated issue if I cannot get over it.

For reference:

When I install azureml-train-automl in old environment.
pip install azureml-train-automl

There are few error line during installation.
tensorflow 1.10.0 has requirement numpy<=1.14.5,>=1.13.3, but you'll have numpy 1.15.1 which is incompatible.
Cannot uninstall 'PyYAML'. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a partial uninstall.

[kritdikoon]

Closes #38