Jonas Mueller

Closing this issue due to lack of activity. Feel free to reopen if you still have questions!

Looking good! Can you open a PR with these changes so we can move the discussion there? In general, this style of docstring is looking good and we'll want to...

Good question! Yes how about just changing the function call to be just one line.

Can you open a PR with the fixed examples that are runnable via `make doctest`? Our team can then help figure out how to make this run inside our CI...

Yes this catboost virtual-ensembles can be useful, particularly for quantifying model uncertainty which quantile regression does not measure. But we should first add uncertainty-quantification as a bonus functionality for regular...

@valeman conformal calibration is now available in autogluon by setting `calibrate = True` (when used for quantile-regression problems), see our example: https://github.com/awslabs/autogluon/blob/master/examples/tabular/example_quantile_regression.py Implementation is thanks to @taesup-aws and based on...

@gomlfx This functionality is not available in the stable release of autogluon yet (will be in the next stable release). To use it now, you need to install either pip...

There's a lot more work involved in than that. The goal is to create a new IssueManager following the instructions from this guide here: https://docs.cleanlab.ai/master/cleanlab/datalab/guide/custom_issue_manager.html This new IssueManager should specifically...

All label error detection algorithms in Cleanlab are designed to account for imperfect ML models trained on finite data. We are not simply reporting the label issues as those examples...

Currently it's just for standard rectangular bounding box error detection. If you find an effective way to apply it to rotating bounding box object detection, we'd love to hear about...