Dataset Preview
Go to dataset viewer
question (string)question_type (string)question_id (int32)image_id (string)features (array 2D)normalized_boxes (array 2D)answer_type (string)label (sequence)
"What is this photo taken looking through?"
"what is this"
458,752,000
"COCO_train2014_000000458752"
"[[0.0,0.0,0.02306804060935974,0.005900494288653135,0.16498039662837982,0.3215763568878174,0.00882311"(...TRUNCATED)
"[[0.44215837121009827,0.6309473514556885,0.5816012024879456,0.9751682281494141],[0.4869205057621002,"(...TRUNCATED)
"other"
{ "ids": [ 1164 ], "weights": [ 1 ] }
"What position is this man playing?"
"what"
458,752,001
"COCO_train2014_000000458752"
"[[0.0,0.0,0.02306804060935974,0.005900494288653135,0.16498039662837982,0.3215763568878174,0.00882311"(...TRUNCATED)
"[[0.44215837121009827,0.6309473514556885,0.5816012024879456,0.9751682281494141],[0.4869205057621002,"(...TRUNCATED)
"other"
{ "ids": [ 1840, 290 ], "weights": [ 0.3, 1 ] }
"What color is the players shirt?"
"what color is the"
458,752,002
"COCO_train2014_000000458752"
"[[0.0,0.0,0.02306804060935974,0.005900494288653135,0.16498039662837982,0.3215763568878174,0.00882311"(...TRUNCATED)
"[[0.44215837121009827,0.6309473514556885,0.5816012024879456,0.9751682281494141],[0.4869205057621002,"(...TRUNCATED)
"other"
{ "ids": [ 1491 ], "weights": [ 1 ] }
"Is this man a professional baseball player?"
"is this"
458,752,003
"COCO_train2014_000000458752"
"[[0.0,0.0,0.02306804060935974,0.005900494288653135,0.16498039662837982,0.3215763568878174,0.00882311"(...TRUNCATED)
"[[0.44215837121009827,0.6309473514556885,0.5816012024879456,0.9751682281494141],[0.4869205057621002,"(...TRUNCATED)
"yes/no"
{ "ids": [ 1403, 425 ], "weights": [ 0.3, 1 ] }
"What color is the snow?"
"what color is the"
262,146,000
"COCO_train2014_000000262146"
"[[0.5141245126724243,0.0,0.2650524377822876,0.0032181860879063606,0.0,0.0,0.2767711281776428,1.05924"(...TRUNCATED)
"[[0.45951759815216064,0.407429039478302,0.689863383769989,0.5900904536247253],[0.5767551064491272,0."(...TRUNCATED)
"other"
{ "ids": [ 3006 ], "weights": [ 1 ] }
"What is the person doing?"
"what is the person"
262,146,001
"COCO_train2014_000000262146"
"[[0.5141245126724243,0.0,0.2650524377822876,0.0032181860879063606,0.0,0.0,0.2767711281776428,1.05924"(...TRUNCATED)
"[[0.45951759815216064,0.407429039478302,0.689863383769989,0.5900904536247253],[0.5767551064491272,0."(...TRUNCATED)
"other"
{ "ids": [ 1561 ], "weights": [ 1 ] }
"What color is the persons headwear?"
"what color is the"
262,146,002
"COCO_train2014_000000262146"
"[[0.5141245126724243,0.0,0.2650524377822876,0.0032181860879063606,0.0,0.0,0.2767711281776428,1.05924"(...TRUNCATED)
"[[0.45951759815216064,0.407429039478302,0.689863383769989,0.5900904536247253],[0.5767551064491272,0."(...TRUNCATED)
"other"
{ "ids": [ 1227, 1333, 2241 ], "weights": [ 0.3, 1, 0.3 ] }
"What is in the person's hand?"
"what is in the"
524,291,000
"COCO_train2014_000000524291"
"[[0.8848278522491455,1.0095982551574707,0.2627096176147461,0.30632075667381287,0.0,0.491780966520309"(...TRUNCATED)
"[[0.0,0.5595235228538513,0.5337058305740356,0.8161495923995972],[0.39334675669670105,0.4297684431076"(...TRUNCATED)
"other"
{ "ids": [ 1712 ], "weights": [ 1 ] }
"Is the dog waiting?"
"is the"
524,291,001
"COCO_train2014_000000524291"
"[[0.8848278522491455,1.0095982551574707,0.2627096176147461,0.30632075667381287,0.0,0.491780966520309"(...TRUNCATED)
"[[0.0,0.5595235228538513,0.5337058305740356,0.8161495923995972],[0.39334675669670105,0.4297684431076"(...TRUNCATED)
"yes/no"
{ "ids": [ 425 ], "weights": [ 1 ] }
"Is the dog looking at a tennis ball or frisbee?"
"is the"
524,291,002
"COCO_train2014_000000524291"
"[[0.8848278522491455,1.0095982551574707,0.2627096176147461,0.30632075667381287,0.0,0.491780966520309"(...TRUNCATED)
"[[0.0,0.5595235228538513,0.5337058305740356,0.8161495923995972],[0.39334675669670105,0.4297684431076"(...TRUNCATED)
"other"
{ "ids": [ 1712 ], "weights": [ 1 ] }
README.md exists but content is empty. Use the Edit dataset card button to edit it.
Downloads last month
208
Edit dataset card
Evaluate models HF Leaderboard

Models trained or fine-tuned on Graphcore/vqa-lxmert