In Defense of Lazy Visual Grounding for Open-Vocabulary Semantic Segmentation

Benchmark Model Rank Results
open-vocabulary-semantic-segmentation-on-coco-stuff-171LaVG#5mIoU: 23.2
open-vocabulary-semantic-segmentation-on-pascalvoc-20LaVG#17mIoU: 82.5
open-vocabulary-semantic-segmentation-on-ade20k-150LaVG#23mIoU: 15.8
open-vocabulary-semantic-segmentation-on-pascal-context-59LaVG#23mIoU: 34.7