Matthias Feurer
Matthias Feurer
Thank you very much for reporting. @filipbartek do you still plan to pick up your work on #197 again?
It appears that the property `vector` was removed from george shortly after the fork (https://github.com/dfm/george/commit/a503c3df7decf65a53f4c5c0c419c88c8f398433#diff-833395cfe79e402fa3d01855374655b0L76). I assume using the george version from https://github.com/automl/george helps.
Just for clarification, do you think that these should be part of the pipeline tuned by Auto-sklearn or that there should be a standalone mode `AutoSklearnOutlierDetector`? According to the title...
I'm afraid that I won't have the time to implement something here. Also, I think this is somewhat out of scope for Auto-sklearn if the metrics are not in scikit-learn...
Hey @hvarfner, yes, I think this is highly related to my previous issue #1086. I agree that it would be great if SMAC does not decide by itself to stop...
No, sorry, I don't know if this can be disabled or where the number of retries can be increased.
Unfortunately, it is not possible to upload things to the test server at the moment. Therefore, I just transferred the issue to the main OpenML repository.
> We've agreed to not remove faulty datasets, but just deactivate them. I totally agree on that. > Do you think that we should still clean this up anyway? If...
Is there a specific reason to limit the name of features given that the arff format does not require this?
Yeah, I was basically asking about the ASCII limitation.