Tobias Pitters

I added a draft PR where the code at least runs through. I suspect that there is still something wrong in how we handle new operations (so basically we ignore...

Thanks for the bug report and the analysis. Could you help us with a reproducible example?

I see but just any reproducible example with dummy data would help us. Otherwise we would need to reverse engineer the problem which takes a lot of time and might...

Hey, can you pls point us to the dataset? Could it be that there are a lot of features in this dataset?

we need a reproducible example for this, to figure out if this is something we can change. Furthermore improvements to the KernelExplainer are welcome but not on the roadmap of...

@Rajeev-Radhakrish thanks for the PR, on a first non-thorough pass this looks good. I'll need to do a thorough one soon and want to check if the tests cover all...

Thanks for the report, this is possibly related #7722 and #7694. Could you pls provide steps to reproduce this?

@raphaelsty thanks for coming back to this. I assume you are running in streaming mode? That should prevent these errors but it looks like more people than just you have...

Hi @wasup07, with model_output="probability" on a binary classification SHAP explains the probabilities, the thresholds are only applied on these probabilities, so however you adjust the threshold that does not have...

Take this with a grain of salt, this is just my personal understanding: While you technically can overwrite the new_fingerprint with a string, e.g. ```python t = d.add_column("new_column", col_value, new_fingerprint="dummy_fp")...