Show HN: A Layer for robotics dataset quality
75% less training data. 0.5% difference from full-data performance. Source-available coreset selection and dataset auditing for robot learning.
Summary not yet generated for this article.
Original Text
We are exploring how teams can automatically audit data quality, detect problematic demonstrations, and build smaller, high-quality training sets before spending GPU time.
We are looking for design partners to help shape the Calibra.
Comments (4 of ) ⌛