fix label column incorrectly added by feature extractors in RAI Vision dashboard for automl models #2532
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
fix label column incorrectly added by feature extractors in RAI Vision dashboard for automl models
One of our customers was not able to run error analysis when attaching to compute instance due to label column being incorrectly included in the RAI Vision dashboard extracted metadata. They encountered the error:
This seems to be an issue that occurs only for automl models due to the extra "image_details" column which is not skipped in the logic of
extract_features
function.This PR fixes the errors and adds both a unit test for the
extract_features
function and an e2e test for the RAI vision dashboard on the fridge dataset.I was also able to reproduce the error for our AutoML notebook. Note when viewing image metadata we see the extra label column which should not be included:
Checklist