Train Local MLFLow Project conda env missing packages

[jwa5426]

When trying to run the example notebook how-to-use-azureml/track-and-monitor-experiments/using-mlflow/train-projects-local/train-projects-local.ipynb, the code in the final code block (the one that has {“USE_CONDA” : True}) fails with ModuleNotFoundError: No module named ‘sklearn’

It seems the code is trying to use the local conda environment despite specifying it to build one from the MLFlow project.

[shivp950]

I can take a look, thanks for pointing this out. Did you change any of the dependency files or ran the example as is? What version of the azureml-mlflow package are you using?

[jwa5426]

I ran it as-is, using azureml-mlflow 1.26.0

[shivp950]

Could you share the conda yaml? I believe we have an email thread going, feel free to reply there if that's more convenient. We recently updated the package, I don't think it will solve this issue but I'd suggest updating the version anyway.

[kulhanek-CNGROUP]

Hello, please is there any update regarding this? I ran into this issue and I cannot get past it.

[kulhanek-CNGROUP]

Could you maybe share last working environment you ran the notebook in? Especially the versions of mlflow and azureml