Mavs

The problem here is with the TSFreshFeatureExtractor class. When transformed using a np.array it returns a 211x2510 dataframe as expected, but when transformed using a pd.DataFrame, it returns a 1x2510...

I can't debug the error without a full working example. Please, either share the model file or the code to produce that file.

Actually. The problem isn't `spacy` but `ray[tune]`. Running `pip install -e ".[full, nlp]"` fails with `ERROR: Could not find a version that satisfies the requirement ray[tune]>=1.0.0 (from pycaret[full,nlp]) (from versions:...

And what would you show with multiclass classification tasks? Since there is not one positive class... We could show the predicted probabilities of all classes

Hi. Thanks for pointing it out. The issue is fixed in #2768

Can you confirm that this error occurred using the newest rc release `pip install -U --pre pycaret`? And if so, share a reproducible example. Thanks

Can you share which version of pycaret you are using? Run `pycaret.__version__`. Also, I can't debug without a reproducible example. Thanks

We can't fix the issue without code to reproduce it. If you have private data, please reproduce the error on a public dataset and share it here.

This error seems to appear only for the catboost model. I pushed a fix in #2841. Until the merge, try any other model.

Hi. I’m currently on holidays so I can not do any debugging until the 14th. ATOM makes two copies of the dataset (see https://tvdboom.github.io/ATOM/user_guide/data_management/#memory-considerations) so this means that there will...