[ { "name": "Count objects", "files": [ "examples/coco-count.jpg" ], "state": "See the image.", "type": "choice", "question": "How many giraffes are in the image?", "options": "4: null\n3: null\n6: null\n5: null", "gold": "3" }, { "name": "Yes/no about a photo", "files": [ "examples/vqav2-yesno.jpg" ], "state": "See the image.", "type": "noul", "question": "Do the giraffes have wet feet?", "options": "true: Yes\nfalse: No", "gold": "false" }, { "name": "Spatial relation", "files": [ "examples/coco-relation.jpg" ], "state": "See the image.", "type": "choice", "question": "Considering the relative positions of the dining table and the umbrella in the image provided, where is the dining table located with respect to the umbrella?", "options": "below: null\nabove: null", "gold": "below" }, { "name": "Which object is closer", "files": [ "examples/clevr-depth.jpg" ], "state": "See the image.", "type": "choice", "question": "Which of these two objects is closer to the camera?", "options": "the_small_yellow_rubber_sphere: the small yellow rubber sphere\nthe_small_gray_metal_cube: the small gray metal cube", "gold": "the_small_yellow_rubber_sphere" }, { "name": "Count by attribute", "files": [ "examples/clevr-count.jpg" ], "state": "See the image.", "type": "choice", "question": "How many objects are the same color as the matte sphere?", "options": "2: null\n0: null\n3: null\n1: null", "gold": "0" }, { "name": "Read a chart", "files": [ "examples/synth-chart-reason.png" ], "state": "See the figure.", "type": "noul", "question": "In the bottom-left subplot, is the final value of Simulation higher than its initial value?", "options": "true: Yes\nfalse: No", "gold": "false" }, { "name": "Receipt check", "files": [ "examples/cord-kie.jpg" ], "state": "See the image.", "type": "noul", "question": "Does this receipt list an item named \"Dimsum\"?", "options": "true: Yes\nfalse: No", "gold": "false" }, { "name": "Pick a UI element", "files": [ "examples/rico-sca.jpg" ], "state": "See the app screenshot.", "type": "choice", "question": "The red numbered boxes mark elements of this app screen. Command: \"tap the all sports\". Which marked element should be tapped to carry out the command?", "options": "element_3: Element 3\nelement_2: Element 2\nelement_1: Element 1\nelement_4: Element 4", "gold": "element_2" }, { "name": "Read an app screen", "files": [ "examples/rico-screenqa.jpg" ], "state": "See the app screenshot.", "type": "choice", "question": "What is the user name?", "options": "united_states: United States\nfemale: Female\nlaura: Laura\n55: null", "gold": "laura" }, { "name": "Read text in a photo", "files": [ "examples/textvqa-mc.jpg" ], "state": "See the image.", "type": "choice", "question": "Where in canada is showing on poster?", "options": "jimmy_savile: jimmy savile\nvancouver: null\nglobal: null\n2015: null", "gold": "vancouver" }, { "name": "Common-sense about a scene", "files": [ "examples/aokvqa-mc.jpg" ], "state": "See the image.", "type": "choice", "question": "Why is the truck backed up to the building?", "options": "blocking_door: blocking door\nloading: null\nsales: null\nrobbery: null", "gold": "loading" }, { "name": "Match the caption", "files": [ "examples/coco-capmatch.jpg" ], "state": "See the image.", "type": "choice", "question": "Which caption best describes the image?", "options": "a_woman_in_a_field_flying_a_kite: A woman in a field flying a kite.\npeople_walking_next_to_a_busy_ci: people walking next to a busy city street\na_group_of_cooks_work_in_a_small: A group of cooks work in a small kitchen.\nan_elephant_is_carrying_people_o: An elephant is carrying people on it's back as it walks around a tree.", "gold": "an_elephant_is_carrying_people_o" }, { "name": "Verify a spatial statement", "files": [ "examples/vsr-spatial.jpg" ], "state": "See the image.", "type": "noul", "question": "Is the following statement about the image true? \"The person is behind the laptop.\"", "options": "true: True\nfalse: False", "gold": "true" }, { "name": "Object attribute", "files": [ "examples/clevr-attr.jpg" ], "state": "See the image.", "type": "choice", "question": "What material is the large sphere that is the same color as the large rubber block?", "options": "metal: null\nrubber: null", "gold": "metal" }, { "name": "Egg candling (good)", "files": [ "examples/egg-test/egg-test-good-81.jpg" ], "state": "Candling photo: a light is shone through the egg in a dark room.", "type": "choice", "question": "Is this egg good or bad?", "options": "good: Normal, healthy egg\nbad: Defective, spoiled or damaged egg", "gold": "good" }, { "name": "Egg candling (good)", "files": [ "examples/egg-test/egg-test-good-125.jpg" ], "state": "Candling photo: a light is shone through the egg in a dark room.", "type": "choice", "question": "Is this egg good or bad?", "options": "good: Normal, healthy egg\nbad: Defective, spoiled or damaged egg", "gold": "good" }, { "name": "Egg candling (bad)", "files": [ "examples/egg-test/egg-test-bad-78.jpg" ], "state": "Candling photo: a light is shone through the egg in a dark room.", "type": "choice", "question": "Is this egg good or bad?", "options": "good: Normal, healthy egg\nbad: Defective, spoiled or damaged egg", "gold": "bad" }, { "name": "Egg candling (bad)", "files": [ "examples/egg-test/egg-test-bad-107.jpg" ], "state": "Candling photo: a light is shone through the egg in a dark room.", "type": "choice", "question": "Is this egg good or bad?", "options": "good: Normal, healthy egg\nbad: Defective, spoiled or damaged egg", "gold": "bad" } ]