RAI Dashboard error - cannot import name 'logsumexp' from 'sklearn.utils.fixes'

typod 0

Hi,

When trying to create a responsible AI dashboard via a pipeline job I receive the following error:

Execution failed. User process 'python' exited with status code 1. Please check log file 'user_logs/std_log.txt' for error details. Error: Traceback (most recent call last):
File "create_rai_insights.py", line 225, in
main(args)
File "/mnt/azureml/cr/j/fdfafc7ec99147d3b75510c41dc3f28d/exe/wd/_telemetry/_loggerfactory.py", line 143, in wrapper
return func(*args, **kwargs)
File "create_rai_insights.py", line 174, in main
model_estimator = load_mlflow_model(
File "/mnt/azureml/cr/j/fdfafc7ec99147d3b75510c41dc3f28d/exe/wd/rai_component_utilities.py", line 129, in load_mlflow_model
raise ValueError(
ValueError: Unable to load mlflow model from /mnt/azureml/cr/j/fdfafc7ec99147d3b75510c41dc3f28d/cap/data-capability/wd/INPUT_model_input in current environment due to error:
cannot import name 'logsumexp' from 'sklearn.utils.fixes' (/azureml-envs/responsibleai/lib/python3.8/site-packages/sklearn/utils/fixes.py)

This is using the prebuild RAI env. Any suggestions?

DShree 396 Reputation points

2023-08-28T10:35:33.59+00:00
@typod Thanks for the question, this error can occur when there is a version mismatch between the scikit-learn library and the version of the RAI environment.

Upgrade the scikit-learn library to the latest version by running the following command in your RAI environment:

pip install --upgrade scikit-learn
typod 0 Reputation points

2023-08-28T11:03:02.01+00:00

Thanks, I just tried that and got the same error. I thought the point of environments was to get around all this mess :)
Ramr-msft 17,741 Reputation points

2023-09-01T03:44:45.87+00:00

typod Thanks for the details, We have forwarded to the team to check on this and will update on the same.

Share via

RAI Dashboard error - cannot import name 'logsumexp' from 'sklearn.utils.fixes'

Your answer