leaderboard is showing negative values for f1 metric

[yairVanti]

when i run fit on tabular binary classification dataset i get the following leader board

when i run model.report() it looks better , where the metrics are positive:

[pplonski]

Looks like a bug. Thank you for reporting!

[aksy2512]

Hi, I want to work on this issue

[pplonski]

Hi @aksy2512, thanks, it is all yours!

The first step would be to reproduce the issue. Are you able to reproduce it locally with some fake data?

[aksy2512]

Thank you @pplonski for assigning me the issue. I am not able to reproduce the data locally. It would be great if you could suggest me some resources about it

[pplonski]

Thanks @aksy2512 for looking into it.

@yairVanti could you please provide example data and code to reproduce the issue locally?

[yairVanti]

dataset is attached, target is the last column. electrical_grid.csv

just train AutoML with these params : ("total_time_limit": 120, "ml_task": "binary_classification", "mode": "Perform", "eval_metric": "f1", "n_jobs": -1, "random_state": 0) and get_leaderboard() on the fitted AutoML instance.

[pplonski]

Thank you @yairVanti!