Datasets:

bhatvineet
/

bopask-test

images listlengths 1 1	depths listlengths 1 1	masks listlengths 0 3	messages listlengths 2 2	question_type stringclasses 6 values	question_subtype stringclasses 7 values
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_10_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_36_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with green heads to the location of the gray whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <point...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads to the right of the gray whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_28_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_10_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the spatula with the slotted orange head to the location of the cooking spoon with green heads and light wooden handles? Your response should be a list of 2D points that show the path from target object to goal object in that order pat...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_36_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_40_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to relocate the gray whisk with a gray handle to the position of the whisk featuring a black handle and yellow wires? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., out...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads and light wooden handles farther from the camera than the whisk with a black handle and yellow wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads closer to the camera than the whisk with a black handle and yellow wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_36_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_10_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the gray whisk to the location of the cooking spoon with green heads and light wooden handles? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordin...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_10_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_40_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with green heads to the location of the whisk with a black handle and yellow wires? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its ...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the spatula with the slotted orange head positioned to the right of the whisk featuring a black handle and yellow wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads to the left of the gray colored whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the spatula with the slotted orange head positioned to the left of the whisk featuring a black handle and yellow wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads and light wooden handles farther from the camera than the gray whisk with a gray handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_obj_000040_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the whisk featuring a black handle and yellow wires, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be bet...	pose	2dbbox
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_11_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_10_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with red heads to the location of the cooking spoon with green heads? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_obj_000036_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the cuboid corners of the gray whisk with a matching handle, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y shoul...	pose	2dbbox
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_10_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_28_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with green heads to the location of the spatula with the slotted orange head? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordi...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_11_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_40_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with red heads to the location of the whisk with a black handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML ...	trajectory	2d
[ "core/handal/images/scene_000008_frame_000980.png" ]	[ "core/handal/depth_maps/scene_000008_frame_000980_depth.png" ]	[ "core/handal/masks/scene_000008_frame_000980_target_40_mask.png", "core/handal/masks/scene_000008_frame_000980_goal_28_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the whisk with a black handle and yellow wires to the location of the spatula with a slotted orange head? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output ...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the spatula with the pink head and wooden handle positioned to the right of the whisk featuring a black handle and red wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads and a light wooden handle positioned farther from the camera compared to the green colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	farther
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the spatula with the pink head and wooden handle farther from the camera than the whisk with the black handle and red wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the green colored colander positioned farther from the camera compared to the green colored whisk with a green handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_39_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the whisk with the black handle and red wires to the location of the light green whisk with a light green handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: [].,...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_10_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_34_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with green heads to the location of the green colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <p...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_obj_000039_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the whisk featuring a black handle and red wires, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be betwee...	pose	2dbbox
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_27_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_39_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to relocate the spatula with a pink head to the position of the whisk featuring a black handle and red wires? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its ...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_obj_000037_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the cuboid corners of the light green whisk with a light green handle, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) a...	pose	2dbbox
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_39_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_34_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to relocate the whisk with a black handle and red wires to the position of the green colored colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordin...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads closer to the camera than the green colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	closer
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the spatula with the pink head to the left of the whisk with the black handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads and wooden handle farther from the camera than the spatula with the pink head? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	farther
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_10_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with green heads to the location of the green whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <poin...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_34_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_27_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to relocate the green colander to the spatula with a pink head and wooden handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <p...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_target_27_mask.png", "core/handal/masks/scene_000001_frame_000725_goal_10_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the spatula with the pink head to the location of the cooking spoon with green heads? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in X...	trajectory	2d
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_obj_000010_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the cuboid corners of the cooking spoon with green heads and light wooden handles, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be betwee...	pose	2dbbox
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the green colored colander positioned further back from the camera compared to the whisk with a black handle and red wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the spatula with the pink head and wooden handle closer to the camera than the whisk with the black handle and red wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with green heads closer to the camera than the spatula with a pink head? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	closer
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[ "core/handal/masks/scene_000001_frame_000725_obj_000027_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the cuboid corners of the spatula featuring a pink head and wooden handle, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,192...	pose	2dbbox
[ "core/handal/images/scene_000001_frame_000725.png" ]	[ "core/handal/depth_maps/scene_000001_frame_000725_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the green colored colander closer to the camera than the whisk with the black handle and red wires? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000006_frame_001074_grasp.png" ]	[ "core/handal/depth_maps/scene_000006_frame_001074_depth.png" ]	[ "core/handal/masks/scene_000006_frame_001074_obj_000039_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the whisk featuring a black handle and red wires? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip: []., out...	grasp	2dplane
[ "core/handal/images/scene_000006_frame_001074_grasp.png" ]	[ "core/handal/depth_maps/scene_000006_frame_001074_depth.png" ]	[ "core/handal/masks/scene_000006_frame_001074_obj_000010_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the cooking spoon with green heads and light wooden handles? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger ti...	grasp	2dplane
[ "core/handal/images/scene_000003_frame_000721_grasp.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000721_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000721_obj_000037_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the light green whisk with a light green handle? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip: []., outp...	grasp	2dplane
[ "core/handal/images/scene_000003_frame_000721_grasp.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000721_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000721_obj_000036_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the gray whisk with a gray handle? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip: []., output its coordin...	grasp	2dplane
[ "core/handal/images/scene_000003_frame_000721_grasp.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000721_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000721_obj_000011_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the cooking spoon with red heads and light wooden handles? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip:...	grasp	2dplane
[ "core/handal/images/scene_000003_frame_000721_grasp.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000721_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000721_obj_000033_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the turquoise colored rounded colander? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip: []., output its co...	grasp	2dplane
[ "core/handal/images/scene_000004_frame_000939_grasp.png" ]	[ "core/handal/depth_maps/scene_000004_frame_000939_depth.png" ]	[ "core/handal/masks/scene_000004_frame_000939_obj_000006_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the hammer with a red and gray handle? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip: []., output its coo...	grasp	2dplane
[ "core/handal/images/scene_000004_frame_000939_grasp.png" ]	[ "core/handal/depth_maps/scene_000004_frame_000939_depth.png" ]	[ "core/handal/masks/scene_000004_frame_000939_obj_000031_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that indicate the grasp plane for the metal skimmers with long handles and distinctive oval mesh shape? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right...	grasp	2dplane
[ "core/handal/images/scene_000004_frame_000939_grasp.png" ]	[ "core/handal/depth_maps/scene_000004_frame_000939_depth.png" ]	[ "core/handal/masks/scene_000004_frame_000939_obj_000017_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that outline the grasp plane for the bright yellow measuring cup? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger tip: []., output its coordinat...	grasp	2dplane
[ "core/handal/images/scene_000004_frame_000939_grasp.png" ]	[ "core/handal/depth_maps/scene_000004_frame_000939_depth.png" ]	[ "core/handal/masks/scene_000004_frame_000939_obj_000010_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat are the five 2D points in image pixel space that establish the grasp plane for the cooking spoon with green heads and light wooden handles? Your response should be as Grasp center: [], Left finger base: [], Right finger base: [], Left finger tip: [], Right finger ...	grasp	2dplane
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_37_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_36_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the light green whisk with a light green handle to the location of the gray whisk with a gray handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_37_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_11_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the green whisk to the location of the cooking spoon with red heads and light wooden handles? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordina...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_11_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to relocate the cooking spoon with red heads to the position of the light green whisk with a light green handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output i...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon further from the camera than the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	farther
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon closer to the camera than the gray colored whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the turquoise colored rounded colander positioned to the left of the gray colored whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_37_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_33_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to reposition the green whisk to the location of the turquoise rounded colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <po...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_obj_000033_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the turquoise colored rounded colander, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be between (0,1440)...	pose	2dbbox
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_33_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to relocate the turquoise rounded colander to the location of the green colored whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML form...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the turquoise colored rounded colander positioned to the right of the gray colored whisk with a gray handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_obj_000036_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the gray colored whisk with a gray handle, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be between (0,14...	pose	2dbbox
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_37_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_16_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the green whisk to the location of the teal cooking spoon? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <points x y>objec...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and light wooden handles closer to the camera than the light green whisk with a light green handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_36_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_33_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the gray colored whisk to the location of the turquoise colored rounded colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML f...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000741_target_11_mask.png", "core/handal/masks/scene_000003_frame_000741_goal_36_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with red heads to the location of the gray whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <points ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000741.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000741_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the turquoise colored rounded colander closer to the camera than the gray colored whisk with a gray handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_33_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_11_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the turquoise rounded colander to the location of the cooking spoon with red heads and light wooden handles? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., outp...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and a light wooden handle positioned farther from the camera than the green whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon farther from the camera than the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_16_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the teal colored cooking spoon to the location of the green colored whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_36_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_33_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the gray whisk to the position of the turquoise rounded colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <points x ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and light wooden handles closer to the camera than the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_obj_000036_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the gray colored whisk with a gray handle, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be between (0,14...	pose	2dbbox
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_36_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_16_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the gray whisk to the location of the teal cooking spoon? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <points x y>object...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_11_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_33_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with red heads to the location of the turquoise rounded colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and light wooden handles closer to the camera than the green colored whisk with a green handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal cooking spoon positioned to the left of the turquoise rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_33_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_36_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the turquoise colored rounded colander to the location of the gray colored whisk with a gray handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its c...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_11_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_36_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with red heads to the location of the gray whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <points ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_11_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_16_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the cooking spoon with red heads to the location of the teal colored cooking spoon? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and light wooden handles positioned to the right of the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_obj_000011_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the cooking spoon with red heads and light wooden handles, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should ...	pose	2dbbox
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads positioned to the left of the teal colored cooking spoon? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_16_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_11_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the teal colored cooking spoon to the location of the cooking spoon with red heads and light wooden handles? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., outp...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_16_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_33_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the teal colored cooking spoon to the location of the turquoise colored rounded colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_33_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the turquoise colored rounded colander to the location of the green colored whisk with a green handle? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon farther from the camera than the green colored whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and light wooden handles farther from the camera than the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	depth_relative	farther
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_37_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_33_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the green colored whisk to the location of the turquoise colored rounded colander? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon closer to the camera than the green colored whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_36_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_11_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the gray whisk to the location of the cooking spoon with red heads and light wooden handles? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinat...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon closer to the camera than the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_obj_000033_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the turquoise colored rounded colander, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be between (0,1440)...	pose	2dbbox
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_36_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_37_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the gray whisk to the location of the green whisk? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates in XML format <points x y>object</point...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_target_33_mask.png", "core/handal/masks/scene_000003_frame_000144_goal_16_mask.png" ]	[ { "role": "user", "content": "<image>\nWhat is the 2D pixel path to move the turquoise colored rounded colander to the location of the teal colored cooking spoon? Your response should be a list of 2D points that show the path from target object to goal object in that order path: []., output its coordinates ...	trajectory	2d
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads positioned to the left of the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon positioned to the right of the green colored whisk? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position
[ "core/handal/images/scene_000003_frame_000144.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000144_depth.png" ]	[ "core/handal/masks/scene_000003_frame_000144_obj_000037_mask.png" ]	[ { "role": "user", "content": "<image>\nLocate the cuboid corners of the green whisk with a matching green handle, output its bbox coordinates using JSON format. The x and y coordinates of your output must obey the width and height of the image and thus x should be between (0,1920) and y should be between (0...	pose	2dbbox
[ "core/handal/images/scene_000003_frame_000186.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000186_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the cooking spoon with red heads and light wooden handles closer to the camera than the light green whisk with a light green handle? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "no" } ]	depth_relative	closer
[ "core/handal/images/scene_000003_frame_000186.png" ]	[ "core/handal/depth_maps/scene_000003_frame_000186_depth.png" ]	[]	[ { "role": "user", "content": "<image>\nIs the teal colored cooking spoon with the handle positioned to the right of the turquoise colored rounded colander? Your response should be a single word: yes or no." }, { "role": "assistant", "content": "yes" } ]	spatial_reasoning	relative_position

End of preview. Expand in Data Studio

BOPASK-Test

Human-verified evaluation benchmark for the BOPASK spatial-reasoning VQA dataset.

Contains 934 question-answer pairs across two testsets:

core — BOPASK-Core: three BOP-Challenge families (HANDAL, HOPE, YCB-V).
lab — BOPASK-Lab : an in-the-wild set of "home / lab" scenes.

Contents at a glance

Split	Family	Records	RGB images	Depth maps	Masks
core	handal	251	43	41	138
core	hope	189	50	29	231
core	ycbv	248	48	48	153
lab	home	246	21	12 (⚠)	52
Total		934	162	130	574

Question-type distribution

question_type / subtype	handal	hope	ycbv	home	Total
pose / 2dbbox	39	39	38	39	155
grasp / 2dplane	40	40	40	38	158
spatial_reasoning / relative_position	40	40	40	71	191
trajectory / 2d	40	40	40	48	168
depth_relative / closer	40	—	40	16	96
depth_relative / farther	40	—	40	24	104
object_rearrangement / point_wise	12	30	10	10	62
family total	251	189	248	246	934

Layout

bopask-test/
├── README.md
├── core/                           (BOPASK-Core testset)
│   ├── bopask-test-handal.json
│   ├── bopask-test-hope.json
│   ├── bopask-test-ycbv.json
│   ├── handal/
│   │   ├── images/                 (43 *.png)
│   │   ├── depth_maps/             (41 *_depth.png)
│   │   └── masks/                  (138 *_mask.png)
│   ├── hope/
│   │   └── images/  depth_maps/  masks/
│   └── ycbv/
│       └── images/  depth_maps/  masks/
└── lab/                            (BOPASK-Lab testset)
    ├── bopask-test-home.json
    └── home/
        ├── images/                 (21 *.png)
        ├── depth_maps/             (empty — see caveat above)
        └── masks/                  (52 masks_<scene>_<object>.png)

All paths inside each JSON are relative to this dataset root, e.g. core/handal/images/scene_000008_frame_000980.png.

Quick start

import json
from datasets import load_dataset

# Load one of the configs:
ds = load_dataset("bhatvineet/bopask-test", "core-handal", split="test")
print(ds[0])

# Or load all four families manually:
configs = ["core-handal", "core-hope", "core-ycbv", "lab-home"]
for cfg in configs:
    d = load_dataset("bhatvineet/bopask-test", cfg, split="test")
    print(cfg, len(d))

Loading directly without datasets:

import json
with open("core/bopask-test-handal.json") as f:
    records = json.load(f)

for r in records:
    img_path  = r["images"][0]          # e.g. "core/handal/images/..."
    user_q    = r["messages"][0]["content"]
    gt_answer = r["messages"][1]["content"]

Evaluation protocols

Each record is a single-turn VQA pair with one ground-truth response in messages[1].content. Answer formats are self-describing — the user prompt tells the model the expected output format (e.g. "respond as a list of 2D points…"). Common metrics by type:

question_type	typical metric
pose / 2dbbox	2D IoU
grasp / 2dplane	endpoint L2 / success@τ
trajectory / 2d	trajectory-wise DTW, endpoint error
spatial_reasoning / relative_position	exact match (yes/no)
depth_relative	exact match (closer/farther)
object_rearrangement / point_wise	point-in-mask accuracy

Relationship to the training set

This benchmark was curated and human-verified to be disjoint from the bhatvineet/bopask-train training split. Use this for evaluation only.

Citation

If you use this dataset, please cite the BOPASK paper and the underlying BOP-Challenge object-pose datasets (HANDAL, HOPE, LineMOD, YCB-V).

License

MIT for the QA annotations. The underlying RGB / depth / mask assets inherit the licenses of their source BOP-Challenge datasets (HANDAL, HOPE, YCB-V) and the bopask-home captures.

Downloads last month: 11

Paper for bhatvineet/bopask-test

BOP-ASK: Object-Interaction Reasoning for Vision-Language Models

Paper • 2511.16857 • Published Dec 4, 2025