How can I use a different environment for the components executed in the environment/update the libraries version

Xavier Val 0

I'm trying to create a ML model capable of explaining prediction results using the SHAP library in Azure ML Designer using the following code:

User's image However, when I execute the pipeline, I encounter the following error:

User's image

Upon researching, I discovered that I need to upgrade NumPy to at least version 1.8.0. The default version in the environment used by the designer and pipeline is 1.22: User's image It seems there's no straightforward way to modify the Conda configuration to import the required version, nor have I found a way to change the environment used for the pipeline designer.

Is there any workaround to upgrade the version, or am I missing something regarding how to change the environment?

YutongTie-MSFT 51,696 Reputation points

2024-07-11T00:03:06.7533333+00:00

Hello,

Thanks for reaching out to us, a good way to make a customized environment is you can create a new custom environment and leverage it directly so that you don't need to make change in the curated environment.

Please have a try and let us know how it works.

Regards,

Yutong
Xavier Val 0 Reputation points

2024-07-11T07:40:37.67+00:00

Hello, I have been able to create the environment, but I don't know how to change the environment of the workflow I created.

Share via

How can I use a different environment for the components executed in the environment/update the libraries version

Your answer