Evaluation pipeline run fails with error "Directory 'X' is empty or not exist"

I’m finishing “UiPath AI Fabric Overview” academy course section “Practice - AI Fabric”. I’m following the instructions to train and evaluate the EnglishTextClassification ML Package with the train.csv and test.csv files.

Training succeeds fine but evaluation run with the trained version 2.1 always fails with error “Directory at ‘d5d9e038-53f1-48e7-a44e-61fe59d7a7c2/adc7becb-2915-4a11-b0a1-375e9e5af603’ is empty or not exist”. What could be the reason?

Full stack trace is attached.
log.txt (7.7 KB)


What path are you giving as evaluation dataset while starting the pipeline and where is your test file in datasets?

Thanks for replying. It led me to the obvious solution.

I was 100% sure I uploaded test data in the evalution dataset but, lo and behold, the dataset was empty when I checked. That was really dumb.

I added the .csv file in the dataset and the run completed successfully.

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.