Datasets:
id stringlengths 1 3 | set stringclasses 1
value | env stringclasses 1
value | depth int32 1 1 | question_type stringclasses 30
values | scene_id stringlengths 12 12 | image imagewidth (px) 448 448 | prompt stringlengths 314 1.64k | steps_prompt stringlengths 852 2.51k | answer stringclasses 15
values | program stringlengths 574 2.45k | boxes stringlengths 690 789 |
|---|---|---|---|---|---|---|---|---|---|---|---|
0 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000006 | Follow these steps in order:
S0 = the characters in the image with weapon whose color is not blue.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories:... | Follow these steps in order:
S0 = the characters in the image with weapon whose color is not blue.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories:... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_13", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with weapon whose color is not blue.", "object_ids": [0, 3], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the... | [[604.9107142857143, 613.8392857142857, 763.3928571428571, 794.6428571428571], [42.410714285714285, 843.75, 165.17857142857142, 993.3035714285714], [435.26785714285717, 821.4285714285714, 555.8035714285714, 986.6071428571429], [236.60714285714286, 415.17857142857144, 366.07142857142856, 604.9107142857143], [225.4464285... | |
1 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000033 | Follow these steps in order:
S0 = the characters in the image with a red weapon.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a ... | Follow these steps in order:
S0 = the characters in the image with a red weapon.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a ... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_0", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a red weapon.", "object_ids": [0, 7, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the group have different color... | [[236.60714285714286, 225.44642857142858, 368.30357142857144, 395.0892857142857], [51.339285714285715, 845.9821428571429, 145.08928571428572, 986.6071428571429], [42.410714285714285, 446.42857142857144, 142.85714285714286, 595.9821428571429], [250.0, 51.339285714285715, 359.375, 191.96428571428572], [828.125, 35.714285... | |
2 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000081 | Follow these steps in order:
S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have different color categories ... | Follow these steps in order:
S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have different color categories ... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_31", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible)... | [[625.0, 415.17857142857144, 754.4642857142857, 595.9821428571429], [424.10714285714283, 647.3214285714286, 542.4107142857143, 787.9464285714286], [53.57142857142857, 843.75, 154.01785714285714, 993.3035714285714], [845.9821428571429, 845.9821428571429, 939.7321428571429, 986.6071428571429], [241.07142857142858, 631.69... | |
3 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000042 | Follow these steps in order:
S0 = the characters in the image whose beard and shoes are the same color.
S1 = Do all characters in the group have different color categories for their beard from each other?
Group: S0.
Possible color categories for beard: red, green, blue, yellow, black.
Note: Colors mean color categories... | Follow these steps in order:
S0 = the characters in the image whose beard and shoes are the same color.
S1 = Do all characters in the group have different color categories for their beard from each other?
Group: S0.
Possible color categories for beard: red, green, blue, yellow, black.
Note: Colors mean color categories... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_25", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose beard and shoes are the same color.", "object_ids": [0, 1, 5], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characte... | [[225.44642857142858, 433.0357142857143, 361.60714285714283, 593.75], [622.7678571428571, 845.9821428571429, 741.0714285714286, 986.6071428571429], [823.6607142857143, 812.5, 982.1428571428571, 993.3035714285714], [837.0535714285714, 247.76785714285714, 937.5, 397.32142857142856], [848.2142857142857, 446.42857142857144... | |
4 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000052 | Follow these steps in order:
S0 = the characters in the image with a blue long-sleeved jacket.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: eve... | Follow these steps in order:
S0 = the characters in the image with a blue long-sleeved jacket.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: eve... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_2", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a blue long-sleeved jacket.", "object_ids": [8, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the group have diff... | [[845.9821428571429, 448.6607142857143, 964.2857142857143, 593.75], [227.67857142857142, 613.8392857142857, 375.0, 794.6428571428571], [448.6607142857143, 647.3214285714286, 566.9642857142857, 792.4107142857143], [225.44642857142858, 236.60714285714286, 343.75, 395.0892857142857], [448.6607142857143, 37.94642857142857,... | |
5 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000062 | Follow these steps in order:
S0 = the characters in the image with shoes whose color is not white/gray.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color catego... | Follow these steps in order:
S0 = the characters in the image with shoes whose color is not white/gray.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color catego... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_24", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with shoes whose color is not white/gray.", "object_ids": [0, 1, 2, 4, 5, 6, 7, 8, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 ... | [[647.3214285714286, 448.6607142857143, 741.0714285714286, 589.2857142857143], [406.25, 433.0357142857143, 542.4107142857143, 595.9821428571429], [426.3392857142857, 216.51785714285714, 555.8035714285714, 406.25], [604.9107142857143, 17.857142857142858, 763.3928571428571, 198.66071428571428], [625.0, 613.8392857142857,... | |
6 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000066 | Follow these steps in order:
S0 = the characters in the image whose beard and long-sleeved jacket are the same color.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mea... | Follow these steps in order:
S0 = the characters in the image whose beard and long-sleeved jacket are the same color.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mea... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_30", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose beard and long-sleeved jacket are the same color.", "object_ids": [5, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do a... | [[629.4642857142857, 250.0, 765.625, 395.0892857142857], [241.07142857142858, 843.75, 341.51785714285717, 993.3035714285714], [625.0, 613.8392857142857, 754.4642857142857, 803.5714285714286], [26.785714285714285, 821.4285714285714, 158.48214285714286, 991.0714285714286], [448.6607142857143, 448.6607142857143, 542.41071... | |
7 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000054 | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have different color categories... | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have different color categories... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_24", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible... | [[435.26785714285717, 17.857142857142858, 564.7321428571429, 198.66071428571428], [803.5714285714286, 843.75, 957.5892857142857, 993.3035714285714], [647.3214285714286, 250.0, 765.625, 395.0892857142857], [604.9107142857143, 446.42857142857144, 738.8392857142857, 595.9821428571429], [848.2142857142857, 49.1071428571428... | |
8 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000018 | Follow these steps in order:
S0 = the characters in the image with a blue weapon.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a... | Follow these steps in order:
S0 = the characters in the image with a blue weapon.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_11", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a blue weapon.", "object_ids": [0, 6, 8], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the group have different col... | [[625.0, 812.5, 754.4642857142857, 1002.2321428571429], [406.25, 247.76785714285714, 540.1785714285714, 397.32142857142856], [823.6607142857143, 812.5, 953.125, 993.3035714285714], [53.57142857142857, 446.42857142857144, 187.5, 595.9821428571429], [207.58928571428572, 812.5, 366.07142857142856, 993.3035714285714], [629... | |
9 | grid2d_d1 | grid2d | 1 | all_different_part_color | scene_000092 | Follow these steps in order:
S0 = the characters in the image with shoes whose color is not red.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: e... | Follow these steps in order:
S0 = the characters in the image with shoes whose color is not red.
S1 = Do all characters in the group have different color categories for their shoes from each other?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: e... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_20", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with shoes whose color is not red.", "object_ids": [0, 2, 3, 4, 5, 6, 7, 8, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do al... | [[450.89285714285717, 446.42857142857144, 584.8214285714286, 595.9821428571429], [252.23214285714286, 446.42857142857144, 352.67857142857144, 595.9821428571429], [629.4642857142857, 51.339285714285715, 758.9285714285714, 191.96428571428572], [31.25, 446.42857142857144, 171.875, 604.9107142857143], [430.80357142857144, ... | |
10 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000034 | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every... | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_18", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with long-sleeved jacket whose color is not blue.", "object_ids": [0, 3, 5], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all c... | [[406.25, 247.76785714285714, 540.1785714285714, 397.32142857142856], [426.3392857142857, 17.857142857142858, 555.8035714285714, 198.66071428571428], [232.14285714285714, 446.42857142857144, 375.0, 595.9821428571429], [830.3571428571429, 647.3214285714286, 939.7321428571429, 787.9464285714286], [243.30357142857142, 629... | |
11 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000021 | Follow these steps in order:
S0 = the characters in the image with red hair on their head.
S1 = Do all characters in the group have the same color category for their long-sleeved jacket?
Group: S0.
Possible color categories for long-sleeved jacket: red, green, blue, yellow.
Note: Colors mean color categories: every sha... | Follow these steps in order:
S0 = the characters in the image with red hair on their head.
S1 = Do all characters in the group have the same color category for their long-sleeved jacket?
Group: S0.
Possible color categories for long-sleeved jacket: red, green, blue, yellow.
Note: Colors mean color categories: every sha... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_2", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with red hair on their head.", "object_ids": [3, 5], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the group have the same... | [[845.9821428571429, 845.9821428571429, 955.3571428571429, 986.6071428571429], [53.57142857142857, 645.0892857142857, 187.5, 794.6428571428571], [647.3214285714286, 448.6607142857143, 765.625, 593.75], [448.6607142857143, 37.94642857142857, 542.4107142857143, 191.96428571428572], [53.57142857142857, 49.107142857142854,... | |
12 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000082 | Follow these steps in order:
S0 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have the same color category for their shoes?... | Follow these steps in order:
S0 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have the same color category for their shoes?... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_24", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).", "object_i... | [[42.410714285714285, 49.107142857142854, 158.48214285714286, 198.66071428571428], [241.07142857142858, 843.75, 352.67857142857144, 1002.2321428571429], [450.89285714285717, 49.107142857142854, 551.3392857142857, 198.66071428571428], [207.58928571428572, 446.42857142857144, 361.60714285714283, 595.9821428571429], [832.... | |
13 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000036 | Follow these steps in order:
S0 = the characters in the image whose hair on their head and shoes are the same color.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: ... | Follow these steps in order:
S0 = the characters in the image whose hair on their head and shoes are the same color.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: ... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_38", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose hair on their head and shoes are the same color.", "object_ids": [5, 6, 7], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do... | [[832.5892857142857, 812.5, 962.0535714285714, 993.3035714285714], [207.58928571428572, 812.5, 366.07142857142856, 993.3035714285714], [37.94642857142857, 17.857142857142858, 167.41071428571428, 207.58928571428572], [647.3214285714286, 633.9285714285714, 763.3928571428571, 787.9464285714286], [424.10714285714283, 433.0... | |
14 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000020 | Follow these steps in order:
S0 = the characters in the image with weapon whose color is not blue.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a c... | Follow these steps in order:
S0 = the characters in the image with weapon whose color is not blue.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a c... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_15", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with weapon whose color is not blue.", "object_ids": [1, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the... | [[638.3928571428571, 446.42857142857144, 738.8392857142857, 595.9821428571429], [8.928571428571429, 216.51785714285714, 167.41071428571428, 397.32142857142856], [42.410714285714285, 433.0357142857143, 145.08928571428572, 595.9821428571429], [439.73214285714283, 49.107142857142854, 540.1785714285714, 198.66071428571428]... | |
15 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000000 | Follow these steps in order:
S0 = the characters in the image with a black beard.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a color counts as th... | Follow these steps in order:
S0 = the characters in the image with a black beard.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a color counts as th... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_14", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a black beard.", "object_ids": [1, 2], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the group have the same color c... | [[232.14285714285714, 647.3214285714286, 361.60714285714283, 787.9464285714286], [232.14285714285714, 446.42857142857144, 370.5357142857143, 604.9107142857143], [430.80357142857144, 435.26785714285717, 564.7321428571429, 589.2857142857143], [832.5892857142857, 216.51785714285714, 962.0535714285714, 397.32142857142856],... | |
16 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000037 | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket and weapon are the same color.
S1 = Do all characters in the group have the same color category for their long-sleeved jacket?
Group: S0.
Possible color categories for long-sleeved jacket: red, green, blue, yellow.
Note: Colors mean... | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket and weapon are the same color.
S1 = Do all characters in the group have the same color category for their long-sleeved jacket?
Group: S0.
Possible color categories for long-sleeved jacket: red, green, blue, yellow.
Note: Colors mean... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_34", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose long-sleeved jacket and weapon are the same color.", "object_ids": [7, 9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do ... | [[241.07142857142858, 645.0892857142857, 363.8392857142857, 794.6428571428571], [832.5892857142857, 622.7678571428571, 964.2857142857143, 792.4107142857143], [424.10714285714283, 424.10714285714283, 555.8035714285714, 593.75], [241.07142857142858, 49.107142857142854, 341.51785714285717, 198.66071428571428], [53.5714285... | |
17 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000027 | Follow these steps in order:
S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have the same color category for their shoes?... | Follow these steps in order:
S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = Do all characters in the group have the same color category for their shoes?... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_30", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).", "object_i... | [[250.0, 51.339285714285715, 343.75, 191.96428571428572], [640.625, 629.4642857142857, 750.0, 794.6428571428571], [33.482142857142854, 247.76785714285714, 171.875, 397.32142857142856], [832.5892857142857, 622.7678571428571, 964.2857142857143, 792.4107142857143], [638.3928571428571, 49.107142857142854, 738.8392857142857... | |
18 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000045 | Follow these steps in order:
S0 = the characters in the image with weapon whose color is not blue.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a c... | Follow these steps in order:
S0 = the characters in the image with weapon whose color is not blue.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a c... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_14", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with weapon whose color is not blue.", "object_ids": [0, 6], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the... | [[629.4642857142857, 845.9821428571429, 765.625, 991.0714285714286], [638.3928571428571, 234.375, 741.0714285714286, 397.32142857142856], [53.57142857142857, 627.2321428571429, 154.01785714285714, 794.6428571428571], [845.9821428571429, 51.339285714285715, 939.7321428571429, 191.96428571428572], [426.3392857142857, 216... | |
19 | grid2d_d1 | grid2d | 1 | all_same_part_color | scene_000039 | Follow these steps in order:
S0 = the characters in the image with yellow hair on their head.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a color ... | Follow these steps in order:
S0 = the characters in the image with yellow hair on their head.
S1 = Do all characters in the group have the same color category for their shoes?
Group: S0.
Possible color categories for shoes: red, green, blue, white/gray, black.
Note: Colors mean color categories: every shade of a color ... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_3", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with yellow hair on their head.", "object_ids": [1, 2, 6], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = Do all characters in the group have th... | [[622.7678571428571, 631.6964285714286, 741.0714285714286, 792.4107142857143], [406.25, 613.8392857142857, 564.7321428571429, 794.6428571428571], [433.0357142857143, 37.94642857142857, 542.4107142857143, 191.96428571428572], [236.60714285714286, 17.857142857142858, 366.07142857142856, 207.58928571428572], [828.125, 430... | |
20 | grid2d_d1 | grid2d | 1 | binding | scene_000090 | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket and shoes are the same color.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray... | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket and shoes are the same color.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_24", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose long-sleeved jacket and shoes are the same color.", "object_ids": [5], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = What ar... | [[631.6964285714286, 37.94642857142857, 741.0714285714286, 191.96428571428572], [37.94642857142857, 49.107142857142854, 154.01785714285714, 207.58928571428572], [625.0, 216.51785714285714, 783.4821428571429, 397.32142857142856], [225.44642857142858, 435.26785714285717, 343.75, 593.75], [241.07142857142858, 645.08928571... | |
21 | grid2d_d1 | grid2d | 1 | binding | scene_000072 | Follow these steps in order:
S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = What are the shoes of S0?
Note: Colors mean color categories: e... | Follow these steps in order:
S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = What are the shoes of S0?
Note: Colors mean color categories: e... | "C." | {"steps": [{"step_ref": "S0", "node_id": "L1_33", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible)... | [[620.5357142857143, 35.714285714285715, 758.9285714285714, 198.66071428571428], [42.410714285714285, 828.125, 151.78571428571428, 993.3035714285714], [241.07142857142858, 229.91071428571428, 357.14285714285717, 406.25], [828.125, 49.107142857142854, 982.1428571428571, 198.66071428571428], [33.482142857142854, 645.0892... | |
22 | grid2d_d1 | grid2d | 1 | binding | scene_000003 | Follow these steps in order:
S0 = the characters in the image with a yellow beard.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray.
Options:
A. black
B. blue
C. gree... | Follow these steps in order:
S0 = the characters in the image with a yellow beard.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray.
Options:
A. black
B. blue
C. gree... | "E." | {"steps": [{"step_ref": "S0", "node_id": "L1_7", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a yellow beard.", "object_ids": [3], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = What are the shoes of S0?\nNote: Colors mean color cate... | [[225.44642857142858, 446.42857142857144, 341.51785714285717, 595.9821428571429], [638.3928571428571, 49.107142857142854, 754.4642857142857, 207.58928571428572], [29.017857142857142, 415.17857142857144, 187.5, 595.9821428571429], [622.7678571428571, 415.17857142857144, 763.3928571428571, 595.9821428571429], [51.3392857... | |
23 | grid2d_d1 | grid2d | 1 | binding | scene_000035 | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = What is the long-sleeved jacket of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, whi... | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = What is the long-sleeved jacket of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, whi... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_26", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with long-sleeved jacket whose color is not blue.", "object_ids": [7], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = What is the lo... | [[435.26785714285717, 613.8392857142857, 564.7321428571429, 794.6428571428571], [29.017857142857142, 613.8392857142857, 158.48214285714286, 794.6428571428571], [839.2857142857143, 232.14285714285714, 948.6607142857143, 397.32142857142856], [439.73214285714283, 843.75, 555.8035714285714, 993.3035714285714], [845.9821428... | |
24 | grid2d_d1 | grid2d | 1 | binding | scene_000046 | Follow these steps in order:
S0 = the characters in the image whose hair on their head and long-sleeved jacket are the same color.
S1 = What is the beard of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color... | Follow these steps in order:
S0 = the characters in the image whose hair on their head and long-sleeved jacket are the same color.
S1 = What is the beard of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_33", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose hair on their head and long-sleeved jacket are the same color.", "object_ids": [8], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": ... | [[26.785714285714285, 424.10714285714283, 158.48214285714286, 593.75], [33.482142857142854, 49.107142857142854, 171.875, 207.58928571428572], [845.9821428571429, 51.339285714285715, 939.7321428571429, 191.96428571428572], [430.80357142857144, 247.76785714285714, 569.1964285714286, 397.32142857142856], [406.25, 17.85714... | |
25 | grid2d_d1 | grid2d | 1 | binding | scene_000079 | Follow these steps in order:
S0 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = What is the hair on the head of S0?
Note: Colors mean color categories: ever... | Follow these steps in order:
S0 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = What is the hair on the head of S0?
Note: Colors mean color categories: ever... | "D." | {"steps": [{"step_ref": "S0", "node_id": "L1_20", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).", "object_i... | [[439.73214285714283, 645.0892857142857, 540.1785714285714, 794.6428571428571], [435.26785714285717, 812.5, 564.7321428571429, 993.3035714285714], [823.6607142857143, 49.107142857142854, 966.5178571428571, 207.58928571428572], [647.3214285714286, 35.714285714285715, 750.0, 198.66071428571428], [42.410714285714285, 446.... | |
26 | grid2d_d1 | grid2d | 1 | binding | scene_000037 | Follow these steps in order:
S0 = the characters in the image where beard, hair on their head, long-sleeved jacket, weapon, and shoes all have different colors from each other.
S1 = What is the beard of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are bot... | Follow these steps in order:
S0 = the characters in the image where beard, hair on their head, long-sleeved jacket, weapon, and shoes all have different colors from each other.
S1 = What is the beard of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are bot... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_36", "filter": "filter_by_all_different_char", "text": "S0 = the characters in the image where beard, hair on their head, long-sleeved jacket, weapon, and shoes all have different colors from each other.", "object_ids": [2], "output_kind": "bboxes"}, {"step_ref": "S1", "node... | [[241.07142857142858, 645.0892857142857, 363.8392857142857, 794.6428571428571], [832.5892857142857, 622.7678571428571, 964.2857142857143, 792.4107142857143], [424.10714285714283, 424.10714285714283, 555.8035714285714, 593.75], [241.07142857142858, 49.107142857142854, 341.51785714285717, 198.66071428571428], [53.5714285... | |
27 | grid2d_d1 | grid2d | 1 | binding | scene_000010 | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = What is the beard of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray.
Optio... | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = What is the beard of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray.
Optio... | "C." | {"steps": [{"step_ref": "S0", "node_id": "L1_14", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with long-sleeved jacket whose color is not blue.", "object_ids": [3], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = What is the be... | [[227.67857142857142, 17.857142857142858, 386.1607142857143, 198.66071428571428], [406.25, 17.857142857142858, 564.7321428571429, 198.66071428571428], [26.785714285714285, 250.0, 145.08928571428572, 395.0892857142857], [821.4285714285714, 424.10714285714283, 953.125, 593.75], [252.23214285714286, 232.14285714285714, 35... | |
28 | grid2d_d1 | grid2d | 1 | binding | scene_000024 | Follow these steps in order:
S0 = the characters in the image with black hair on their head.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray.
Options:
A. black
B. bl... | Follow these steps in order:
S0 = the characters in the image with black hair on their head.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray.
Options:
A. black
B. bl... | "E." | {"steps": [{"step_ref": "S0", "node_id": "L1_5", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with black hair on their head.", "object_ids": [8], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = What are the shoes of S0?\nNote: Colors mean ... | [[638.3928571428571, 247.76785714285714, 738.8392857142857, 397.32142857142856], [250.0, 845.9821428571429, 343.75, 986.6071428571429], [29.017857142857142, 415.17857142857144, 187.5, 595.9821428571429], [647.3214285714286, 647.3214285714286, 741.0714285714286, 787.9464285714286], [53.57142857142857, 247.76785714285714... | |
29 | grid2d_d1 | grid2d | 1 | binding | scene_000048 | Follow these steps in order:
S0 = the characters in the image whose hair on their head and weapon are the same color.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray... | Follow these steps in order:
S0 = the characters in the image whose hair on their head and weapon are the same color.
S1 = What are the shoes of S0?
Note: Colors mean color categories: every shade of a color counts as that color (light blue and dark blue are both blue), and white and gray count as one color, white/gray... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_33", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose hair on their head and weapon are the same color.", "object_ids": [2], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "Q_ANSWER", "filter": null, "text": "S1 = What ar... | [[8.928571428571429, 232.14285714285714, 162.94642857142858, 397.32142857142856], [439.73214285714283, 645.0892857142857, 540.1785714285714, 794.6428571428571], [426.3392857142857, 216.51785714285714, 555.8035714285714, 406.25], [625.0, 17.857142857142858, 783.4821428571429, 198.66071428571428], [821.4285714285714, 845... | |
30 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000047 | Follow these steps in order:
S0 = the characters in the image with hair on their head whose color is not black.
S1 = the characters in the image with hair on their head whose color is not green.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
... | Follow these steps in order:
S0 = the characters in the image with hair on their head whose color is not black.
S1 = the characters in the image with hair on their head whose color is not green.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_22", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with hair on their head whose color is not black.", "object_ids": [1], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_16", "filter": "filter_by_part_value_inequality", "t... | [[439.73214285714283, 627.2321428571429, 540.1785714285714, 794.6428571428571], [26.785714285714285, 633.9285714285714, 162.94642857142858, 792.4107142857143], [604.9107142857143, 49.107142857142854, 738.8392857142857, 198.66071428571428], [622.7678571428571, 830.3571428571429, 758.9285714285714, 991.0714285714286], [6... | |
31 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000097 | Follow these steps in order:
S0 = the characters in the image whose hair on their head and long-sleeved jacket are the same color.
S1 = the characters in the image whose long-sleeved jacket and weapon are the same color.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one c... | Follow these steps in order:
S0 = the characters in the image whose hair on their head and long-sleeved jacket are the same color.
S1 = the characters in the image whose long-sleeved jacket and weapon are the same color.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one c... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_33", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose hair on their head and long-sleeved jacket are the same color.", "object_ids": [1], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_34", "filter": "filter_by_within... | [[37.94642857142857, 415.17857142857144, 167.41071428571428, 595.9821428571429], [448.6607142857143, 435.26785714285717, 542.4107142857143, 589.2857142857143], [232.14285714285714, 250.0, 368.30357142857144, 395.0892857142857], [207.58928571428572, 830.3571428571429, 343.75, 993.3035714285714], [647.3214285714286, 647.... | |
32 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000074 | Follow these steps in order:
S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image with a green long-sleeved jacket.
S2 = Are S0 an... | Follow these steps in order:
S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image with a green long-sleeved jacket.
S2 = Are S0 an... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_29", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).", "object_i... | [[430.80357142857144, 247.76785714285714, 584.8214285714286, 397.32142857142856], [232.14285714285714, 448.6607142857143, 361.60714285714283, 589.2857142857143], [633.9285714285714, 33.482142857142854, 750.0, 207.58928571428572], [250.0, 647.3214285714286, 343.75, 787.9464285714286], [848.2142857142857, 645.08928571428... | |
33 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000037 | Follow these steps in order:
S0 = the characters in the image whose hair on their head and shoes are the same color.
S1 = the characters in the image where beard, hair on their head, long-sleeved jacket, weapon, and shoes all have different colors from each other.
S2 = Are S0 and S1 close to each other on the grid (sho... | Follow these steps in order:
S0 = the characters in the image whose hair on their head and shoes are the same color.
S1 = the characters in the image where beard, hair on their head, long-sleeved jacket, weapon, and shoes all have different colors from each other.
S2 = Are S0 and S1 close to each other on the grid (sho... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_35", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose hair on their head and shoes are the same color.", "object_ids": [7], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_36", "filter": "filter_by_all_different_char",... | [[241.07142857142858, 645.0892857142857, 363.8392857142857, 794.6428571428571], [832.5892857142857, 622.7678571428571, 964.2857142857143, 792.4107142857143], [424.10714285714283, 424.10714285714283, 555.8035714285714, 593.75], [241.07142857142858, 49.107142857142854, 341.51785714285717, 198.66071428571428], [53.5714285... | |
34 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000072 | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = the characters in the image with hair on their head whose color is not green.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
... | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = the characters in the image with hair on their head whose color is not green.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_22", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with long-sleeved jacket whose color is not blue.", "object_ids": [2], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_26", "filter": "filter_by_part_value_inequality", "t... | [[620.5357142857143, 35.714285714285715, 758.9285714285714, 198.66071428571428], [42.410714285714285, 828.125, 151.78571428571428, 993.3035714285714], [241.07142857142858, 229.91071428571428, 357.14285714285717, 406.25], [828.125, 49.107142857142854, 982.1428571428571, 198.66071428571428], [33.482142857142854, 645.0892... | |
35 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000073 | Follow these steps in order:
S0 = the characters in the image with a green long-sleeved jacket.
S1 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S2 = Are S0 an... | Follow these steps in order:
S0 = the characters in the image with a green long-sleeved jacket.
S1 = the characters in the image whose beard color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S2 = Are S0 an... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_8", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a green long-sleeved jacket.", "object_ids": [8], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_26", "filter": "filter_by_unique_attribute", "text": "S1 = the characters in the... | [[625.0, 51.339285714285715, 758.9285714285714, 191.96428571428572], [241.07142857142858, 446.42857142857144, 341.51785714285717, 595.9821428571429], [51.339285714285715, 51.339285714285715, 160.71428571428572, 191.96428571428572], [42.410714285714285, 645.0892857142857, 154.01785714285714, 803.5714285714286], [24.5535... | |
36 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000071 | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = the characters in the image whose beard and shoes are the same color.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
Note: Co... | Follow these steps in order:
S0 = the characters in the image with long-sleeved jacket whose color is not blue.
S1 = the characters in the image whose beard and shoes are the same color.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
Note: Co... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_23", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with long-sleeved jacket whose color is not blue.", "object_ids": [1], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_31", "filter": "filter_by_within_char_same_parts", "... | [[250.0, 633.9285714285714, 343.75, 787.9464285714286], [250.0, 448.6607142857143, 359.375, 589.2857142857143], [633.9285714285714, 415.17857142857144, 763.3928571428571, 604.9107142857143], [424.10714285714283, 622.7678571428571, 555.8035714285714, 792.4107142857143], [37.94642857142857, 821.4285714285714, 169.6428571... | |
37 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000096 | Follow these steps in order:
S0 = the characters in the image whose weapon color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image with a red long-sleeved jacket.
S2 = Are S0 and... | Follow these steps in order:
S0 = the characters in the image whose weapon color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image with a red long-sleeved jacket.
S2 = Are S0 and... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_35", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose weapon color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).", "object_... | [[450.89285714285717, 247.76785714285714, 551.3392857142857, 397.32142857142856], [647.3214285714286, 435.26785714285717, 741.0714285714286, 589.2857142857143], [42.410714285714285, 631.6964285714286, 145.08928571428572, 794.6428571428571], [234.375, 845.9821428571429, 343.75, 986.6071428571429], [629.4642857142857, 83... | |
38 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000029 | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket and shoes are the same color.
S1 = the characters in the image with a red long-sleeved jacket.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
Note: Colors... | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket and shoes are the same color.
S1 = the characters in the image with a red long-sleeved jacket.
S2 = Are S0 and S1 close to each other on the grid (shortest distance apart them smaller than one character's width/height)?
Note: Colors... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_29", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose long-sleeved jacket and shoes are the same color.", "object_ids": [1], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_8", "filter": "filter_by_part_value", "text":... | [[625.0, 17.857142857142858, 765.625, 198.66071428571428], [448.6607142857143, 35.714285714285715, 566.9642857142857, 198.66071428571428], [42.410714285714285, 428.57142857142856, 154.01785714285714, 604.9107142857143], [29.017857142857142, 17.857142857142858, 169.64285714285714, 198.66071428571428], [625.0, 247.767857... | |
39 | grid2d_d1 | grid2d | 1 | character_closeness | scene_000058 | Follow these steps in order:
S0 = the characters in the image with beard whose color is not blue.
S1 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S2 = Are S0 ... | Follow these steps in order:
S0 = the characters in the image with beard whose color is not blue.
S1 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S2 = Are S0 ... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_19", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with beard whose color is not blue.", "object_ids": [1], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_20", "filter": "filter_by_unique_attribute", "text": "S1 = the cha... | [[629.4642857142857, 236.60714285714286, 765.625, 390.625], [629.4642857142857, 830.3571428571429, 765.625, 991.0714285714286], [51.339285714285715, 35.714285714285715, 145.08928571428572, 191.96428571428572], [625.0, 17.857142857142858, 754.4642857142857, 198.66071428571428], [37.94642857142857, 424.10714285714283, 16... | |
40 | grid2d_d1 | grid2d | 1 | closest_character | scene_000073 | Follow these steps in order:
S0 = the characters in the image with hair on their head whose color is not green.
S1 = the characters in the image with a green long-sleeved jacket.
S2 = the characters in the image whose long-sleeved jacket and shoes are the same color.
S3 = the characters in the image with green hair on ... | Follow these steps in order:
S0 = the characters in the image with hair on their head whose color is not green.
S1 = the characters in the image with a green long-sleeved jacket.
S2 = the characters in the image whose long-sleeved jacket and shoes are the same color.
S3 = the characters in the image with green hair on ... | "C." | {"steps": [{"step_ref": "S0", "node_id": "L1_24", "filter": "filter_by_part_value_inequality", "text": "S0 = the characters in the image with hair on their head whose color is not green.", "object_ids": [6], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_8", "filter": "filter_by_part_value", "text": "S1 = ... | [[625.0, 51.339285714285715, 758.9285714285714, 191.96428571428572], [241.07142857142858, 446.42857142857144, 341.51785714285717, 595.9821428571429], [51.339285714285715, 51.339285714285715, 160.71428571428572, 191.96428571428572], [42.410714285714285, 645.0892857142857, 154.01785714285714, 803.5714285714286], [24.5535... | |
41 | grid2d_d1 | grid2d | 1 | closest_character | scene_000020 | Follow these steps in order:
S0 = the characters in the image with a green long-sleeved jacket.
S1 = the characters in the image with black hair on their head.
S2 = the characters in the image with a green beard.
S3 = the characters in the image with a blue beard.
S4 = the characters in the image whose shoes color is u... | Follow these steps in order:
S0 = the characters in the image with a green long-sleeved jacket.
S1 = the characters in the image with black hair on their head.
S2 = the characters in the image with a green beard.
S3 = the characters in the image with a blue beard.
S4 = the characters in the image whose shoes color is u... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_8", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a green long-sleeved jacket.", "object_ids": [7], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_9", "filter": "filter_by_part_value", "text": "S1 = the characters in the image ... | [[638.3928571428571, 446.42857142857144, 738.8392857142857, 595.9821428571429], [8.928571428571429, 216.51785714285714, 167.41071428571428, 397.32142857142856], [42.410714285714285, 433.0357142857143, 145.08928571428572, 595.9821428571429], [439.73214285714283, 49.107142857142854, 540.1785714285714, 198.66071428571428]... | |
42 | grid2d_d1 | grid2d | 1 | closest_character | scene_000030 | Follow these steps in order:
S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image whose long-sleeved jacket and shoes... | Follow these steps in order:
S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image whose long-sleeved jacket and shoes... | "D." | {"steps": [{"step_ref": "S0", "node_id": "L1_25", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose hair on their head color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible)... | [[241.07142857142858, 33.482142857142854, 341.51785714285717, 198.66071428571428], [241.07142857142858, 247.76785714285714, 341.51785714285717, 397.32142857142856], [830.3571428571429, 51.339285714285715, 939.7321428571429, 191.96428571428572], [631.6964285714286, 821.4285714285714, 754.4642857142857, 986.6071428571429... | |
43 | grid2d_d1 | grid2d | 1 | closest_character | scene_000034 | Follow these steps in order:
S0 = the characters in the image with a blue beard.
S1 = the characters in the image with a red weapon.
S2 = the characters in the image with green hair on their head.
S3 = the characters in the image with yellow hair on their head.
S4 = the characters in the image with red shoes.
S5 = Amon... | Follow these steps in order:
S0 = the characters in the image with a blue beard.
S1 = the characters in the image with a red weapon.
S2 = the characters in the image with green hair on their head.
S3 = the characters in the image with yellow hair on their head.
S4 = the characters in the image with red shoes.
S5 = Amon... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_3", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a blue beard.", "object_ids": [4], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_7", "filter": "filter_by_part_value", "text": "S1 = the characters in the image with a red weap... | [[406.25, 247.76785714285714, 540.1785714285714, 397.32142857142856], [426.3392857142857, 17.857142857142858, 555.8035714285714, 198.66071428571428], [232.14285714285714, 446.42857142857144, 375.0, 595.9821428571429], [830.3571428571429, 647.3214285714286, 939.7321428571429, 787.9464285714286], [243.30357142857142, 629... | |
44 | grid2d_d1 | grid2d | 1 | closest_character | scene_000021 | Follow these steps in order:
S0 = the characters in the image whose beard and shoes are the same color.
S1 = the characters in the image whose beard and weapon are the same color.
S2 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a ... | Follow these steps in order:
S0 = the characters in the image whose beard and shoes are the same color.
S1 = the characters in the image whose beard and weapon are the same color.
S2 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a ... | "C." | {"steps": [{"step_ref": "S0", "node_id": "L1_23", "filter": "filter_by_within_char_same_parts", "text": "S0 = the characters in the image whose beard and shoes are the same color.", "object_ids": [7], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_26", "filter": "filter_by_within_char_same_parts", "text": ... | [[845.9821428571429, 845.9821428571429, 955.3571428571429, 986.6071428571429], [53.57142857142857, 645.0892857142857, 187.5, 794.6428571428571], [647.3214285714286, 448.6607142857143, 765.625, 593.75], [448.6607142857143, 37.94642857142857, 542.4107142857143, 191.96428571428572], [53.57142857142857, 49.107142857142854,... | |
45 | grid2d_d1 | grid2d | 1 | closest_character | scene_000031 | Follow these steps in order:
S0 = the characters in the image with a green beard.
S1 = the characters in the image with a yellow beard.
S2 = the characters in the image with hair on their head whose color is not red.
S3 = the characters in the image with a yellow long-sleeved jacket.
S4 = the characters in the image wi... | Follow these steps in order:
S0 = the characters in the image with a green beard.
S1 = the characters in the image with a yellow beard.
S2 = the characters in the image with hair on their head whose color is not red.
S3 = the characters in the image with a yellow long-sleeved jacket.
S4 = the characters in the image wi... | "A." | {"steps": [{"step_ref": "S0", "node_id": "L1_7", "filter": "filter_by_part_value", "text": "S0 = the characters in the image with a green beard.", "object_ids": [9], "output_kind": "bboxes"}, {"step_ref": "S1", "node_id": "L1_1", "filter": "filter_by_part_value", "text": "S1 = the characters in the image with a yellow ... | [[225.44642857142858, 448.6607142857143, 343.75, 593.75], [250.0, 35.714285714285715, 343.75, 191.96428571428572], [33.482142857142854, 830.3571428571429, 187.5, 993.3035714285714], [604.9107142857143, 17.857142857142858, 763.3928571428571, 198.66071428571428], [604.9107142857143, 446.42857142857144, 758.9285714285714,... | |
46 | grid2d_d1 | grid2d | 1 | closest_character | scene_000010 | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image whose shoes color is unique among ... | Follow these steps in order:
S0 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image whose shoes color is unique among ... | "D." | {"steps": [{"step_ref": "S0", "node_id": "L1_23", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose long-sleeved jacket color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible... | [[227.67857142857142, 17.857142857142858, 386.1607142857143, 198.66071428571428], [406.25, 17.857142857142858, 564.7321428571429, 198.66071428571428], [26.785714285714285, 250.0, 145.08928571428572, 395.0892857142857], [821.4285714285714, 424.10714285714283, 953.125, 593.75], [252.23214285714286, 232.14285714285714, 35... | |
47 | grid2d_d1 | grid2d | 1 | closest_character | scene_000053 | Follow these steps in order:
S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image with long-sleeved jacket whose color is not gree... | Follow these steps in order:
S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).
S1 = the characters in the image with long-sleeved jacket whose color is not gree... | "B." | {"steps": [{"step_ref": "S0", "node_id": "L1_26", "filter": "filter_by_unique_attribute", "text": "S0 = the characters in the image whose shoes color is unique among the characters in that set (a color that is unique among a group: compared only with the characters in the group that have that part visible).", "object_i... | [[832.5892857142857, 812.5, 962.0535714285714, 993.3035714285714], [37.94642857142857, 216.51785714285714, 167.41071428571428, 406.25], [439.73214285714283, 446.42857142857144, 540.1785714285714, 595.9821428571429], [42.410714285714285, 428.57142857142856, 154.01785714285714, 604.9107142857143], [424.10714285714283, 64... |
End of preview. Expand in Data Studio
CoVerse bench
2,176 questions about images, for testing how well vision-language models follow multi-step visual reasoning.
The images come from four sources: 2D grids of characters (grid2d), rendered 3D characters (chars3d), ProcTHOR
house interiors (procthor3d) and real indoor scans (ca1m3d). Each question is built from a small program. Depth 1
asks one thing; depth 3 chains several steps. There is one split per environment and depth, 12 in total, named like
grid2d_d1 or ca1m3d_d3.
from datasets import load_dataset
ds = load_dataset("Gigglingface/coverse-bench", split="procthor3d_d2")
Running a model on it
The code is at https://github.com/oshapio/coverse-platform. It works with any model behind an OpenAI-compatible API (vLLM, OpenRouter, OpenAI, ...).
git clone https://github.com/oshapio/coverse-platform.git
cd coverse-platform
pip install -e .
# add your model to models.yaml, then:
python -m coverse_bench.run --set grid2d_d1 --model vllm --mode answer
python -m coverse_bench.score --set grid2d_d1 --preds outputs/grid2d_d1/vllm_answer.jsonl
--mode steps asks for boxes at every step as well as the answer. The repo README explains the modes and the
scoring, and has a loop over all 12 splits.
Columns
id: the question's id within its split. Answers are matched to questions by this id.image: the scene.prompt: the question as asked in answer mode (just give the answer).steps_prompt: the question with its program written out step by step. In this mode the model gives boxes for every step and then the answer.answer: the correct answer, as JSON (a letter, a number or an ordered list).program: the steps, with the ids of the objects each step picks out, as JSON.boxes: the box of every object in the scene, on a 0 to 1000 grid over the image, as JSON.question_type,depth,env,set,scene_id.
- Downloads last month
- 70