Scoring and sorting all samples during textcat teach

A post was merged into an existing topic: Prodigy Active Learning prefer_uncertain mechanism