RefCOCO: Results as Published

RefCOCO, RefCOCO+, and RefCOCOg test referring expressions on COCO images: given a phrase, the model should localize that object. RefCOCO and RefCOCO+ were collected in a two-player game, RefCOCO+ forbids location words, and RefCOCOg was written separately without that game. The paper counts 142,209 expressions for 50,000 objects in 19,994 images for RefCOCO, and 141,564 expressions for 49,856 objects in 19,992 images for RefCOCO+. For comprehension, a predicted box counts when its intersection-over-union with the reference box is above 0.5, and that rate is averaged over images.

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.