Datasets:
images images listlengths 1 1 | image_group_id stringlengths 6 11 | qa listlengths 1 67 |
|---|---|---|
coco:75910 | [
{
"view": "refcoco",
"inputs": "upper person in brown",
"choices_list": [
"[555, 458, 571, 478]",
"[0, 7, 115, 244]",
"[237, 2, 495, 494]",
"[516, 49, 797, 530]",
"[90, 636, 481, 975]",
"[537, 446, 559, 487]",
"[58, 0, 206, 153]",
"[522, 520, 978, 1000]"
... | |
coco:301603 | [
{
"view": "refcocog",
"inputs": "a train chair across from two men looking at a laptop",
"choices_list": [
"[30, 647, 161, 754]",
"[370, 372, 428, 503]",
"[425, 5, 989, 778]",
"[0, 135, 252, 875]",
"[137, 498, 703, 986]",
"[350, 341, 402, 421]",
"[2, 748, 210, 9... | |
coco:459853 | [
{
"view": "refcocog",
"inputs": "Wooden chair with man in glasses sitting on it.",
"choices_list": [
"[10, 363, 142, 981]",
"[564, 0, 992, 387]",
"[252, 116, 782, 979]",
"[179, 434, 329, 979]",
"[248, 835, 344, 916]",
"[510, 847, 557, 930]"
],
"labels": 3,
... | |
coco:486304 | [
{
"view": "refcocog",
"inputs": "A horse to the left of another horse",
"choices_list": [
"[597, 632, 764, 989]",
"[264, 758, 432, 986]",
"[754, 613, 929, 1000]",
"[201, 49, 219, 76]",
"[263, 773, 297, 892]",
"[306, 733, 331, 780]",
"[36, 779, 54, 850]",
"... | |
coco:360473 | [
{
"view": "refcocog",
"inputs": "Lady in a white blazer to the right of a table and a woman in a white sweater.",
"choices_list": [
"[458, 920, 485, 954]",
"[388, 616, 549, 897]",
"[351, 885, 385, 925]",
"[576, 157, 948, 984]",
"[806, 392, 1000, 965]",
"[232, 843, 673... | |
coco:226705 | [
{
"view": "refcocog",
"inputs": "A vehicle with wood paneling",
"choices_list": [
"[260, 150, 309, 182]",
"[771, 180, 897, 188]",
"[815, 201, 876, 223]",
"[1, 497, 296, 1000]",
"[230, 149, 291, 201]",
"[312, 199, 699, 825]",
"[407, 112, 602, 637]",
"[126, ... | |
[{"src":"https://datasets-server.huggingface.co/assets/tasksource/refcoco-grounding/--/{dataset_git_(...TRUNCATED) | coco:59394 | [{"view":"refcoco","inputs":"middle guy looking up holding phoen","choices_list":["[0, 542, 280, 703(...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/tasksource/refcoco-grounding/--/{dataset_git_(...TRUNCATED) | coco:399399 | [{"view":"refcocog","inputs":"A blue and white neck tie.","choices_list":["[15, 231, 142, 740]","[47(...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/tasksource/refcoco-grounding/--/{dataset_git_(...TRUNCATED) | coco:258559 | [{"view":"refcocog","inputs":"backside of a purple 3 wheeled motorcycle","choices_list":["[282, 226,(...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/tasksource/refcoco-grounding/--/{dataset_git_(...TRUNCATED) | coco:477156 | [{"view":"refcocog","inputs":"The large sofa.","choices_list":["[3, 286, 425, 849]","[452, 59, 595, (...TRUNCATED) |
refcoco-grounding
Human referring expressions from RefCOCO, RefCOCO+ and RefCOCOg, joined to COCO2014.
One shared image cache and mirror; views retain their native question identities. The union of all held-out image IDs is reserved from training across variants; overlapping validation/test images stay in test only. Original split names and per-image Flickr licenses are recorded without excluding noncommercial or unresolved sources. Original images are redistributed unchanged; optional pixel augmentations remain subject to each image's original terms. Only native COCO object boxes supply options; no invented distractors.
Original data: jxu124/refcoco, jxu124/refcocoplus, jxu124/refcocog, Untitled111/refcoco-train2014-assets. Repackaged as parquet for tasksource by scripts/repackage_dataset/grounding.py.
Images are stored once with grouped qa records: train: 100,000 questions / 8,526 image groups, validation: 24,527 questions / 2,566 image groups, test: 30,903 questions / 3,980 image groups.
Conversion provenance: source revisions, hashes, configuration and exclusions.
Original referring-expression release · Coarse recast audit.
- Downloads last month
- -