Code crashes while trying to reproduce the result

[rsarbhangia]

I'm trying to reproduce the bechmark results. I've installed the packages as per requirements.

Error: Problem Set - coffee-quality-data-cqi Question: 4
Getting AttributeError: 'Dataframe' object has no attribute 'dtype'.

Command: python scripts/test.py benchmarks/kaggle --model gpt-35-turbo --endpoint openai

I was assuming this should have been registered as a crash error since the code is not running properly. Any help would be much appreciated.

[ultmaster]

Would you try pandas < 1.5, scikit-learn < 1.4?

You can also use this docker: https://hub.docker.com/r/ultmaster/dseval

I'll mark documentation of used library version and updates of the used libraries as a pending work item.

[rsarbhangia]

Tried. Now I've pandas = 1.4.4 and scikit-learn = 1.3.2
Now getting NotImplementedError: cannot aling with a higher dimensional NDFrame

[ultmaster]

Would you show the whole traceback?

My guess is numpy needs to be downgraded too...