www.omnilabel.org

About

Task The next generation of perception should understand complex free-form object descriptions, rather than a fixed set of categories. To accelerate this vision, we propose a novel & challenging benchmark. Checkout our task description and paper for more details