| {"id": 0, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the grapes in the blue plate, then put the banana in the pink plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the grapes in the blue plate, then put the banana in the pink plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp grapes', 'move grasped object to blue plate', 'release', 'grasp banana', 'move grasped object to pink plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 1, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the grapes in the blue plate, then put the banana in the pink plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the grapes in the blue plate, then put the banana in the pink plate. There may be missing subtasks, or wrong subtasks. Plan: ['move grasped object to blue plate', 'release', 'grasp banana', 'move grasped object to pink plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> missing subtasks: one or several subtasks are missing </category>"}]} | |
| {"id": 2, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to remove the screwdriver out of the middle drawer and place it on the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: remove the screwdriver out of the middle drawer and place it on the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the middle drawer', 'move grasped object out', 'release', 'grasp screwdriver', 'move grasped object to table', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 3, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to remove the screwdriver out of the middle drawer and place it on the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: remove the screwdriver out of the middle drawer and place it on the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp screwdriver', 'move grasped object to table', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> missing subtasks: one or several subtasks are missing </category>"}]} | |
| {"id": 4, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to remove the apple out of the bottom drawer and place it on top of the drawer shelf, then close the drawer. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: remove the apple out of the bottom drawer and place it on top of the drawer shelf, then close the drawer. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the middle drawer', 'move grasped object out', 'release', 'grasp apple', 'move grasped object to top of the drawer shelf', 'release', 'grasp handle of the middle drawer', 'move grasped object to drawer closed position', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 5, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to remove the apple out of the bottom drawer and place it on top of the drawer shelf, then close the drawer. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: remove the apple out of the bottom drawer and place it on top of the drawer shelf, then close the drawer. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the middle drawer', 'move grasped object out', 'release', 'grasp apple', 'move grasped object to top of the drawer shelf', 'release', 'grasp handle of the middle drawer', 'move grasped object to drawer open position', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object </category>"}]} | |
| {"id": 6, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to open all the drawers. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: open all the drawers. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the top drawer', 'move grasped object out', 'release', 'grasp handle of the middle drawer', 'move grasped object out', 'release', 'grasp handle of the bottom drawer', 'move grasped object out', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 7, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to open all the drawers. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: open all the drawers. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the top drawer', 'move grasped object out', 'release', 'grasp handle of the bottom drawer', 'move grasped object out', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> missing subtasks: one or several subtasks are missing </category>"}]} | |
| {"id": 8, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to close all the drawers that appear open. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: close all the drawers that appear open. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the top drawer', 'move grasped object to drawer closed position', 'release', 'grasp handle of the middle drawer', 'move grasped object to drawer closed position', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 9, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to close all the drawers that appear open. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: close all the drawers that appear open. There may be missing subtasks, or wrong subtasks. Plan: ['grasp handle of the top drawer', 'move grasped object to drawer closed position', 'release', 'grasp handle of the bottom drawer', 'move grasped object to drawer closed position', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object manipulated: one or several of the subtasks manipulate the wrong object </category>"}]} | |
| {"id": 10, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack everything and put them back on the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack everything and put them back on the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp can', 'move grasped object to top of pink cube', 'release', 'grasp frog toy', 'move grasped object to top of can', 'release', 'grasp frog toy', 'move grasped object to table', 'release', 'grasp can', 'move grasped object to table', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 11, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack everything and put them back on the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack everything and put them back on the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp can', 'move grasped object to top of pink cube', 'release', 'grasp frog toy', 'move grasped object to top of can', 'release', 'grasp frog toy', 'move grasped object to table', 'release', 'grasp can', 'move grasped object to top of frog toy', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object </category>"}]} | |
| {"id": 12, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack the frog toy and put it next to the stack but not onto. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack the frog toy and put it next to the stack but not onto. There may be missing subtasks, or wrong subtasks. Plan: ['grasp can', 'move grasped object to top of pink cube', 'release', 'grasp frog toy', 'move grasped object to top of can', 'release', 'grasp frog toy', 'move grasped object next to but not onto the stack', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 13, "image": ["records/real_put_item_in_drawer+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack the frog toy and put it next to the stack but not onto. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the can onto the pink cube, then stack the frog toy on top of this stack, then unstack the frog toy and put it next to the stack but not onto. There may be missing subtasks, or wrong subtasks. Plan: ['grasp can', 'move grasped object to top of pink cube', 'release', 'grasp frog toy', 'move grasped object next to but not onto the stack', 'release', 'grasp frog toy', 'move grasped object to top of can', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong order: one or several subtasks are not in the right order, breaking causal dependencies </category>"}]} | |
| {"id": 14, "image": ["records/real_stack_cup+0/episode0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack a yellow cup onto a pink cup onto a navy cup. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack a yellow cup onto a pink cup onto a navy cup. There may be missing subtasks, or wrong subtasks. Plan: ['grasp yellow cup', 'move grasped object to pink cup', 'release', 'grasp pink cup', 'move grasped object to navy cup', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 15, "image": ["records/real_stack_cup+0/episode0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack a yellow cup onto a pink cup onto a navy cup. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack a yellow cup onto a pink cup onto a navy cup. There may be missing subtasks, or wrong subtasks. Plan: ['grasp yellow cup', 'move grasped object to pink cup', 'release', 'grasp pink cup', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> missing subtasks: one or several subtasks are missing </category>"}]} | |
| {"id": 16, "image": ["records/real_stack_cups+0/0/19_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack red cup onto yellow cup onto black cup. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack red cup onto yellow cup onto black cup. There may be missing subtasks, or wrong subtasks. Plan: ['grasp red cup', 'move grasped object to yellow cup', 'release', 'grasp yellow cup', 'move grasped object to black cup', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 17, "image": ["records/real_stack_cups+0/0/19_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack yellow cup onto red cup onto black cup. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack yellow cup onto red cup onto black cup. There may be missing subtasks, or wrong subtasks. Plan: ['grasp red cup', 'move grasped object to yellow cup', 'release', 'grasp yellow cup', 'move grasped object to black cup', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object </category>"}]} | |
| {"id": 18, "image": ["records/real_pick_item_in_box+0/episode2/2_img_viewpoint_left.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to store the food in the container. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: store the food in the container. There may be missing subtasks, or wrong subtasks. Plan: ['grasp croissant', 'move grasped object to box', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 19, "image": ["records/real_pick_item_in_box+0/episode2/2_img_viewpoint_left.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to place the food on the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: place the food on the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp croissant', 'move grasped object to box', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object </category>"}]} | |
| {"id": 20, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the banana, the grapes, the strawberry, yellow peach, yellow lemon, green pear in the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the banana, the grapes, the strawberry, yellow peach, yellow lemon, green pear in the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp banana', 'move grasped object to blue plate', 'release', 'grasp grapes', 'move grasped object to blue plate', 'release', 'grasp strawberry', 'move grasped object to blue plate', 'release', 'grasp yellow peach', 'move grasped object to blue plate', 'release', 'grasp yellow lemon', 'move grasped object to blue plate', 'release', 'grasp green pear', 'move grasped object to blue plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 21, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the banana, the grapes, the strawberry, yellow peach, yellow lemon, green pear in the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the banana, the grapes, the strawberry, yellow peach, yellow lemon, green pear in the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp banana', 'move grasped object to blue plate', 'release', 'grasp grapes', 'move grasped object to blue plate', 'release', 'grasp raspberry', 'move grasped object to blue plate', 'release', 'grasp yellow peach', 'move grasped object to blue plate', 'release', 'grasp yellow lemon', 'move grasped object to blue plate', 'release', 'grasp green pear', 'move grasped object to blue plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object manipulated: one or several of the subtasks manipulate the wrong object </category>"}]} | |
| {"id": 22, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the banana, the grapes in the pink plate, then the strawberry, yellow peach, yellow lemon, green pear in the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the banana, the grapes in the pink plate, then the strawberry, yellow peach, yellow lemon, green pear in the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp banana', 'move grasped object to pink plate', 'release', 'grasp grapes', 'move grasped object to pink plate', 'release', 'grasp raspberry', 'move grasped object to blue plate', 'release', 'grasp yellow peach', 'move grasped object to blue plate', 'release', 'grasp yellow lemon', 'move grasped object to blue plate', 'release', 'grasp green pear', 'move grasped object to blue plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 23, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the banana, the grapes in the pink plate, then the strawberry, yellow peach, yellow lemon, green pear in the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the banana, the grapes in the pink plate, then the strawberry, yellow peach, yellow lemon, green pear in the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp banana', 'move grasped object to pink plate', 'release', 'grasp grapes', 'move grasped object to pink plate', 'release', 'grasp raspberry', 'move grasped object to blue plate', 'release', 'grasp yellow peach', 'move grasped object to blue plate', 'release', 'grasp yellow lemon', 'move grasped object to blue plate', 'release', 'grasp green pear', 'move grasped object to pink plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object </category>"}]} | |
| {"id": 24, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the yellow peach in the blue plate, then to the pink plate, then again to the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the yellow peach in the blue plate, then to the pink plate, then again to the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp yellow peach', 'move grasped object to blue plate', 'release', 'grasp yellow peach', 'move grasped object to pink plate', 'release', 'grasp yellow peach', 'move grasped object to blue plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 25, "image": ["records/real_put_fruits_in_plate+0/0/2_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to put the yellow peach in the blue plate, then to the pink plate, then again to the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: put the yellow peach in the blue plate, then to the pink plate, then again to the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp yellow peach', 'move grasped object to blue plate', 'release', 'grasp yellow peach', 'move grasped object to pink plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> missing subtasks: one or several subtasks are missing </category>"}]} | |
| {"id": 26, "image": ["records/real_put_fruits_in_plate+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the blue plate on top of the pink plate, then put the strawberry in the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the blue plate on top of the pink plate, then put the strawberry in the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp blue plate', 'move grasped object to pink plate', 'release', 'grasp strawberry', 'move grasped object to blue plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 27, "image": ["records/real_put_fruits_in_plate+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the blue plate on top of the pink plate, then put the strawberry in the blue plate. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the blue plate on top of the pink plate, then put the strawberry in the blue plate. There may be missing subtasks, or wrong subtasks. Plan: ['grasp strawberry', 'move grasped object to blue plate', 'release', 'grasp blue plate', 'move grasped object to pink plate', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> wrong order: one or several subtasks are not in the right order, breaking causal dependencies </category>"}]} | |
| {"id": 28, "image": ["records/real_put_fruits_in_plate+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the blue plate on top of the pink plate, then put the strawberry, and yellow lemon and banana in the blue plate, then move back the banana to the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the blue plate on top of the pink plate, then put the strawberry, and yellow lemon and banana in the blue plate, then move back the banana to the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp blue plate', 'move grasped object to pink plate', 'release', 'grasp strawberry', 'move grasped object to blue plate', 'release', 'grasp yellow lemon', 'move grasped object to blue plate', 'release', 'grasp banana', 'move grasped object to blue plate', 'release', 'grasp banana', 'move grasped object out', 'move grasped object to table', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> True </answer> <category> success </category>"}]} | |
| {"id": 29, "image": ["records/real_put_fruits_in_plate+0/0/0_img_viewpoint_0.png"], "height_list": [256], "width_list": [256], "conversations": [{"from": "human", "value": "Image of the scene before action: <image> \n\n You are observing a manipulation task performed by a robot equipped with a gripper. The robot's high-level objective is to stack the blue plate on top of the pink plate, then put the strawberry, and yellow lemon and banana in the blue plate, then move back the banana to the table. The task is divided into multiple subtasks. Verify if the robot's plan is correct (=True), or incorrect (=False) for the task: stack the blue plate on top of the pink plate, then put the strawberry, and yellow lemon and banana in the blue plate, then move back the banana to the table. There may be missing subtasks, or wrong subtasks. Plan: ['grasp blue plate', 'move grasped object to pink plate', 'release', 'grasp strawberry', 'move grasped object to blue plate', 'release', 'grasp yellow lemon', 'release', 'move grasped object to blue plate', 'release', 'grasp banana', 'move grasped object to blue plate', 'release', 'grasp banana', 'move grasped object out', 'move grasped object to table', 'release'] Answer with a boolean between True or False enclosed within <answer> </answer> tags. Finally, categorize the success/failure category by enclosing it within <category> </category> tags among: ['success', 'missing subtasks: one or several subtasks are missing', 'wrong object manipulated: one or several of the subtasks manipulate the wrong object', 'wrong object state or placement: one or several of the subtasks select the wrong target/location or state for the correct object', 'wrong order: one or several subtasks are not in the right order, breaking causal dependencies', 'contradictory subtasks: some subtasks conflict with each other']. E.g <answer> boolean </answer> <category> category </category>"}, {"from": "gpt", "value": "<answer> False </answer> <category> contradictory subtasks: some subtasks conflict with each other </category>"}]} | |