[ { "image": "data/prediction/urban_day_rittenhouse/image/000000.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000000.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 154.5, "y1": 188.0, "x2": 328.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000000" }, { "image": "data/prediction/urban_day_rittenhouse/image/000000.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000000.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 154.5, "y1": 188.0, "x2": 328.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000000" }, { "image": "data/prediction/urban_day_rittenhouse/image/000001.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000001.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 153.0, "y1": 189.0, "x2": 326.5, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000001" }, { "image": "data/prediction/urban_day_rittenhouse/image/000001.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000001.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 153.0, "y1": 189.0, "x2": 326.5, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000001" }, { "image": "data/prediction/urban_day_rittenhouse/image/000002.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000002.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 154.0, "y1": 190.0, "x2": 325.0, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000002" }, { "image": "data/prediction/urban_day_rittenhouse/image/000002.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000002.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 154.0, "y1": 190.0, "x2": 325.0, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000002" }, { "image": "data/prediction/urban_day_rittenhouse/image/000003.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000003.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 153.0, "y1": 192.5, "x2": 330.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000003" }, { "image": "data/prediction/urban_day_rittenhouse/image/000003.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000003.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 153.0, "y1": 192.5, "x2": 330.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000003" }, { "image": "data/prediction/urban_day_rittenhouse/image/000004.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000004.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 157.0, "y1": 200.0, "x2": 361.0, "y2": 273.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000004" }, { "image": "data/prediction/urban_day_rittenhouse/image/000004.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000004.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 157.0, "y1": 200.0, "x2": 361.0, "y2": 273.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000004" }, { "image": "data/prediction/urban_day_rittenhouse/image/000005.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000005.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 274.0, "y1": 204.5, "x2": 523.5, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000005" }, { "image": "data/prediction/urban_day_rittenhouse/image/000005.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000005.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 274.0, "y1": 204.5, "x2": 523.5, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000005" }, { "image": "data/prediction/urban_day_rittenhouse/image/000006.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000006.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 194.0, "x2": 16.5, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000006" }, { "image": "data/prediction/urban_day_rittenhouse/image/000006.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000006.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 194.0, "x2": 16.5, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000006" }, { "image": "data/prediction/urban_day_rittenhouse/image/000007.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000007.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 203.5, "x2": 90.0, "y2": 268.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000007" }, { "image": "data/prediction/urban_day_rittenhouse/image/000007.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000007.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 203.5, "x2": 90.0, "y2": 268.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000007" }, { "image": "data/prediction/urban_day_rittenhouse/image/000008.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000008.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 68.0, "y1": 204.0, "x2": 251.0, "y2": 287.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000008" }, { "image": "data/prediction/urban_day_rittenhouse/image/000008.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000008.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 68.0, "y1": 204.0, "x2": 251.0, "y2": 287.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000008" }, { "image": "data/prediction/urban_day_rittenhouse/image/000009.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000009.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 260.5, "y1": 214.0, "x2": 519.0, "y2": 343.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000009" }, { "image": "data/prediction/urban_day_rittenhouse/image/000009.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000009.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 260.5, "y1": 214.0, "x2": 519.0, "y2": 343.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000009" }, { "image": "data/prediction/urban_day_rittenhouse/image/000010.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000010.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 198.0, "x2": 37.0, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000010" }, { "image": "data/prediction/urban_day_rittenhouse/image/000010.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000010.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 198.0, "x2": 37.0, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000010" }, { "image": "data/prediction/urban_day_rittenhouse/image/000011.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000011.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 70.0, "y1": 201.5, "x2": 206.5, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000011" }, { "image": "data/prediction/urban_day_rittenhouse/image/000011.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000011.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 70.0, "y1": 201.5, "x2": 206.5, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000011" }, { "image": "data/prediction/urban_day_rittenhouse/image/000012.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000012.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 295.5, "y1": 210.0, "x2": 478.0, "y2": 313.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000012" }, { "image": "data/prediction/urban_day_rittenhouse/image/000012.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000012.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 295.5, "y1": 210.0, "x2": 478.0, "y2": 313.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000012" }, { "image": "data/prediction/urban_day_rittenhouse/image/000013.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000013.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 171.0, "y1": 196.0, "x2": 250.0, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000013" }, { "image": "data/prediction/urban_day_rittenhouse/image/000013.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000013.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 171.0, "y1": 196.0, "x2": 250.0, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000013" }, { "image": "data/prediction/urban_day_rittenhouse/image/000014.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000014.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 381.5, "y1": 202.5, "x2": 488.5, "y2": 268.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000014" }, { "image": "data/prediction/urban_day_rittenhouse/image/000014.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000014.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 381.5, "y1": 202.5, "x2": 488.5, "y2": 268.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000014" }, { "image": "data/prediction/urban_day_rittenhouse/image/000015.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000015.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 193.0, "x2": 71.5, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000015" }, { "image": "data/prediction/urban_day_rittenhouse/image/000015.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000015.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 193.0, "x2": 71.5, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000015" }, { "image": "data/prediction/urban_day_rittenhouse/image/000016.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000016.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 220.5, "y1": 200.0, "x2": 328.0, "y2": 269.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000016" }, { "image": "data/prediction/urban_day_rittenhouse/image/000016.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000016.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 220.5, "y1": 200.0, "x2": 328.0, "y2": 269.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000016" }, { "image": "data/prediction/urban_day_rittenhouse/image/000017.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000017.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 450.5, "y1": 210.0, "x2": 612.0, "y2": 317.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000017" }, { "image": "data/prediction/urban_day_rittenhouse/image/000017.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000017.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 450.5, "y1": 210.0, "x2": 612.0, "y2": 317.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000017" }, { "image": "data/prediction/urban_day_rittenhouse/image/000018.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000018.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 395.0, "y1": 198.5, "x2": 447.0, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000018" }, { "image": "data/prediction/urban_day_rittenhouse/image/000018.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000018.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 395.0, "y1": 198.5, "x2": 447.0, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000018" }, { "image": "data/prediction/urban_day_rittenhouse/image/000019.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000019.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 414.0, "y1": 199.0, "x2": 471.0, "y2": 242.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000019" }, { "image": "data/prediction/urban_day_rittenhouse/image/000019.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000019.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.0, "y1": 199.0, "x2": 471.0, "y2": 242.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000019" }, { "image": "data/prediction/urban_day_rittenhouse/image/000020.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000020.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 446.5, "y1": 202.0, "x2": 518.0, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000020" }, { "image": "data/prediction/urban_day_rittenhouse/image/000020.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000020.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 446.5, "y1": 202.0, "x2": 518.0, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000020" }, { "image": "data/prediction/urban_day_rittenhouse/image/000021.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000021.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 521.5, "y1": 209.5, "x2": 628.0, "y2": 314.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000021" }, { "image": "data/prediction/urban_day_rittenhouse/image/000021.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000021.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 521.5, "y1": 209.5, "x2": 628.0, "y2": 314.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000021" }, { "image": "data/prediction/urban_day_rittenhouse/image/000022.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000022.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 354.0, "y1": 195.5, "x2": 404.0, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000022" }, { "image": "data/prediction/urban_day_rittenhouse/image/000022.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000022.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 354.0, "y1": 195.5, "x2": 404.0, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000022" }, { "image": "data/prediction/urban_day_rittenhouse/image/000023.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000023.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 357.0, "y1": 196.5, "x2": 426.0, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000023" }, { "image": "data/prediction/urban_day_rittenhouse/image/000023.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000023.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.0, "y1": 196.5, "x2": 426.0, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000023" }, { "image": "data/prediction/urban_day_rittenhouse/image/000024.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000024.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 365.0, "y1": 201.5, "x2": 463.5, "y2": 288.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000024" }, { "image": "data/prediction/urban_day_rittenhouse/image/000024.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000024.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 365.0, "y1": 201.5, "x2": 463.5, "y2": 288.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000024" }, { "image": "data/prediction/urban_day_rittenhouse/image/000025.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000025.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 387.5, "y1": 209.0, "x2": 538.0, "y2": 347.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000025" }, { "image": "data/prediction/urban_day_rittenhouse/image/000025.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000025.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 387.5, "y1": 209.0, "x2": 538.0, "y2": 347.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000025" }, { "image": "data/prediction/urban_day_rittenhouse/image/000026.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000026.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.5, "y1": 214.5, "x2": 635.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000026" }, { "image": "data/prediction/urban_day_rittenhouse/image/000026.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000026.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.5, "y1": 214.5, "x2": 635.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000026" }, { "image": "data/prediction/urban_day_rittenhouse/image/000027.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000027.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.0, "y1": 222.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000027" }, { "image": "data/prediction/urban_day_rittenhouse/image/000027.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000027.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.0, "y1": 222.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000027" }, { "image": "data/prediction/urban_day_rittenhouse/image/000028.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000028.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 175.5, "y1": 184.5, "x2": 229.5, "y2": 203.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000028" }, { "image": "data/prediction/urban_day_rittenhouse/image/000028.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000028.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 175.5, "y1": 184.5, "x2": 229.5, "y2": 203.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000028" }, { "image": "data/prediction/urban_day_rittenhouse/image/000029.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000029.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 151.5, "y1": 188.0, "x2": 212.0, "y2": 209.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000029" }, { "image": "data/prediction/urban_day_rittenhouse/image/000029.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000029.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 151.5, "y1": 188.0, "x2": 212.0, "y2": 209.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000029" }, { "image": "data/prediction/urban_day_rittenhouse/image/000030.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000030.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 116.5, "y1": 187.0, "x2": 187.5, "y2": 212.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000030" }, { "image": "data/prediction/urban_day_rittenhouse/image/000030.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000030.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 116.5, "y1": 187.0, "x2": 187.5, "y2": 212.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000030" }, { "image": "data/prediction/urban_day_rittenhouse/image/000031.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000031.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 442.0, "y1": 200.5, "x2": 503.0, "y2": 253.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000031" }, { "image": "data/prediction/urban_day_rittenhouse/image/000031.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000031.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 442.0, "y1": 200.5, "x2": 503.0, "y2": 253.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000031" }, { "image": "data/prediction/urban_day_rittenhouse/image/000032.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000032.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 490.0, "y1": 201.5, "x2": 574.0, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000032" }, { "image": "data/prediction/urban_day_rittenhouse/image/000032.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000032.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 490.0, "y1": 201.5, "x2": 574.0, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000032" }, { "image": "data/prediction/urban_day_rittenhouse/image/000033.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000033.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 445.5, "y1": 201.5, "x2": 522.5, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000033" }, { "image": "data/prediction/urban_day_rittenhouse/image/000033.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000033.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 445.5, "y1": 201.5, "x2": 522.5, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000033" }, { "image": "data/prediction/urban_day_rittenhouse/image/000034.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000034.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 479.0, "y1": 205.0, "x2": 580.5, "y2": 275.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000034" }, { "image": "data/prediction/urban_day_rittenhouse/image/000034.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000034.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 479.0, "y1": 205.0, "x2": 580.5, "y2": 275.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000034" }, { "image": "data/prediction/urban_day_rittenhouse/image/000035.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000035.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 526.5, "y1": 212.0, "x2": 639.5, "y2": 314.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000035" }, { "image": "data/prediction/urban_day_rittenhouse/image/000035.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000035.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 526.5, "y1": 212.0, "x2": 639.5, "y2": 314.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000035" }, { "image": "data/prediction/urban_day_rittenhouse/image/000036.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000036.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 583.5, "y1": 214.5, "x2": 639.5, "y2": 358.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000036" }, { "image": "data/prediction/urban_day_rittenhouse/image/000036.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000036.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 583.5, "y1": 214.5, "x2": 639.5, "y2": 358.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000036" }, { "image": "data/prediction/urban_day_rittenhouse/image/000037.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000037.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 413.0, "y1": 192.0, "x2": 474.5, "y2": 216.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000037" }, { "image": "data/prediction/urban_day_rittenhouse/image/000037.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000037.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 413.0, "y1": 192.0, "x2": 474.5, "y2": 216.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000037" }, { "image": "data/prediction/urban_day_rittenhouse/image/000038.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000038.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 390.5, "y1": 192.5, "x2": 462.5, "y2": 218.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000038" }, { "image": "data/prediction/urban_day_rittenhouse/image/000038.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000038.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 390.5, "y1": 192.5, "x2": 462.5, "y2": 218.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000038" }, { "image": "data/prediction/urban_day_rittenhouse/image/000039.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000039.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 375.5, "y1": 191.5, "x2": 450.0, "y2": 218.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000039" }, { "image": "data/prediction/urban_day_rittenhouse/image/000039.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000039.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 375.5, "y1": 191.5, "x2": 450.0, "y2": 218.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000039" }, { "image": "data/prediction/urban_day_rittenhouse/image/000040.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000040.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 382.0, "y1": 190.5, "x2": 439.5, "y2": 218.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000040" }, { "image": "data/prediction/urban_day_rittenhouse/image/000040.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000040.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 382.0, "y1": 190.5, "x2": 439.5, "y2": 218.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000040" }, { "image": "data/prediction/urban_day_rittenhouse/image/000041.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000041.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 365.5, "y1": 197.0, "x2": 423.5, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000041" }, { "image": "data/prediction/urban_day_rittenhouse/image/000041.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000041.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 365.5, "y1": 197.0, "x2": 423.5, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000041" }, { "image": "data/prediction/urban_day_rittenhouse/image/000042.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000042.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.0, "y1": 197.5, "x2": 445.5, "y2": 264.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000042" }, { "image": "data/prediction/urban_day_rittenhouse/image/000042.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000042.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.0, "y1": 197.5, "x2": 445.5, "y2": 264.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000042" }, { "image": "data/prediction/urban_day_rittenhouse/image/000043.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000043.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 371.0, "y1": 200.5, "x2": 475.5, "y2": 280.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000043" }, { "image": "data/prediction/urban_day_rittenhouse/image/000043.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000043.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 371.0, "y1": 200.5, "x2": 475.5, "y2": 280.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000043" }, { "image": "data/prediction/urban_day_rittenhouse/image/000044.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000044.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 385.5, "y1": 203.0, "x2": 505.0, "y2": 297.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000044" }, { "image": "data/prediction/urban_day_rittenhouse/image/000044.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000044.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 385.5, "y1": 203.0, "x2": 505.0, "y2": 297.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000044" }, { "image": "data/prediction/urban_day_rittenhouse/image/000045.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000045.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 442.5, "y1": 194.5, "x2": 527.5, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000045" }, { "image": "data/prediction/urban_day_rittenhouse/image/000045.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000045.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 442.5, "y1": 194.5, "x2": 527.5, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000045" }, { "image": "data/prediction/urban_day_rittenhouse/image/000046.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000046.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 475.5, "y1": 195.0, "x2": 592.0, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000046" }, { "image": "data/prediction/urban_day_rittenhouse/image/000046.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000046.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 475.5, "y1": 195.0, "x2": 592.0, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000046" }, { "image": "data/prediction/urban_day_rittenhouse/image/000047.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000047.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 505.5, "y1": 198.0, "x2": 639.5, "y2": 292.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000047" }, { "image": "data/prediction/urban_day_rittenhouse/image/000047.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000047.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 505.5, "y1": 198.0, "x2": 639.5, "y2": 292.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000047" }, { "image": "data/prediction/urban_day_rittenhouse/image/000048.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000048.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 520.0, "y1": 206.0, "x2": 639.5, "y2": 308.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000048" }, { "image": "data/prediction/urban_day_rittenhouse/image/000048.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000048.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 520.0, "y1": 206.0, "x2": 639.5, "y2": 308.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000048" }, { "image": "data/prediction/urban_day_rittenhouse/image/000049.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000049.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 54.5, "y1": 181.0, "x2": 143.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000049" }, { "image": "data/prediction/urban_day_rittenhouse/image/000049.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000049.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 54.5, "y1": 181.0, "x2": 143.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000049" }, { "image": "data/prediction/urban_day_rittenhouse/image/000050.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000050.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 32.0, "y1": 175.5, "x2": 128.0, "y2": 210.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000050" }, { "image": "data/prediction/urban_day_rittenhouse/image/000050.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000050.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 32.0, "y1": 175.5, "x2": 128.0, "y2": 210.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000050" }, { "image": "data/prediction/urban_day_rittenhouse/image/000051.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000051.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 461.0, "y1": 197.0, "x2": 542.5, "y2": 253.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000051" }, { "image": "data/prediction/urban_day_rittenhouse/image/000051.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000051.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 461.0, "y1": 197.0, "x2": 542.5, "y2": 253.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000051" }, { "image": "data/prediction/urban_day_rittenhouse/image/000052.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000052.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 475.0, "y1": 203.0, "x2": 571.0, "y2": 264.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000052" }, { "image": "data/prediction/urban_day_rittenhouse/image/000052.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000052.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 475.0, "y1": 203.0, "x2": 571.0, "y2": 264.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000052" }, { "image": "data/prediction/urban_day_rittenhouse/image/000053.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000053.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 482.0, "y1": 208.5, "x2": 588.5, "y2": 277.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000053" }, { "image": "data/prediction/urban_day_rittenhouse/image/000053.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000053.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 482.0, "y1": 208.5, "x2": 588.5, "y2": 277.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000053" }, { "image": "data/prediction/urban_day_rittenhouse/image/000054.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000054.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 483.5, "y1": 209.5, "x2": 585.5, "y2": 277.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000054" }, { "image": "data/prediction/urban_day_rittenhouse/image/000054.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000054.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 483.5, "y1": 209.5, "x2": 585.5, "y2": 277.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000054" }, { "image": "data/prediction/urban_day_rittenhouse/image/000055.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000055.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 485.0, "y1": 210.5, "x2": 586.5, "y2": 278.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000055" }, { "image": "data/prediction/urban_day_rittenhouse/image/000055.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000055.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 485.0, "y1": 210.5, "x2": 586.5, "y2": 278.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000055" }, { "image": "data/prediction/urban_day_rittenhouse/image/000056.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000056.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 496.0, "y1": 209.5, "x2": 589.5, "y2": 278.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000056" }, { "image": "data/prediction/urban_day_rittenhouse/image/000056.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000056.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 496.0, "y1": 209.5, "x2": 589.5, "y2": 278.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000056" }, { "image": "data/prediction/urban_day_rittenhouse/image/000057.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000057.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 495.5, "y1": 212.0, "x2": 597.0, "y2": 282.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000057" }, { "image": "data/prediction/urban_day_rittenhouse/image/000057.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000057.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 495.5, "y1": 212.0, "x2": 597.0, "y2": 282.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000057" }, { "image": "data/prediction/urban_day_rittenhouse/image/000058.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000058.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 503.5, "y1": 211.5, "x2": 607.0, "y2": 284.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000058" }, { "image": "data/prediction/urban_day_rittenhouse/image/000058.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000058.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 503.5, "y1": 211.5, "x2": 607.0, "y2": 284.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000058" }, { "image": "data/prediction/urban_day_rittenhouse/image/000059.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000059.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 504.5, "y1": 211.5, "x2": 622.0, "y2": 289.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000059" }, { "image": "data/prediction/urban_day_rittenhouse/image/000059.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000059.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 504.5, "y1": 211.5, "x2": 622.0, "y2": 289.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000059" }, { "image": "data/prediction/urban_day_rittenhouse/image/000060.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000060.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 521.0, "y1": 216.0, "x2": 635.5, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000060" }, { "image": "data/prediction/urban_day_rittenhouse/image/000060.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000060.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 521.0, "y1": 216.0, "x2": 635.5, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000060" }, { "image": "data/prediction/urban_day_rittenhouse/image/000061.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000061.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 527.5, "y1": 214.0, "x2": 639.5, "y2": 298.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000061" }, { "image": "data/prediction/urban_day_rittenhouse/image/000061.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000061.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 527.5, "y1": 214.0, "x2": 639.5, "y2": 298.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000061" }, { "image": "data/prediction/urban_day_rittenhouse/image/000062.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000062.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 532.5, "y1": 217.5, "x2": 639.5, "y2": 308.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000062" }, { "image": "data/prediction/urban_day_rittenhouse/image/000062.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000062.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 532.5, "y1": 217.5, "x2": 639.5, "y2": 308.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000062" }, { "image": "data/prediction/urban_day_rittenhouse/image/000063.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000063.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 425.5, "y1": 226.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000063" }, { "image": "data/prediction/urban_day_rittenhouse/image/000063.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000063.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 425.5, "y1": 226.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000063" }, { "image": "data/prediction/urban_day_rittenhouse/image/000064.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000064.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.0, "y1": 228.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000064" }, { "image": "data/prediction/urban_day_rittenhouse/image/000064.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000064.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.0, "y1": 228.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000064" }, { "image": "data/prediction/urban_day_rittenhouse/image/000065.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000065.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 432.0, "y1": 226.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000065" }, { "image": "data/prediction/urban_day_rittenhouse/image/000065.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000065.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 432.0, "y1": 226.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000065" }, { "image": "data/prediction/urban_day_rittenhouse/image/000066.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000066.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 433.0, "y1": 226.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000066" }, { "image": "data/prediction/urban_day_rittenhouse/image/000066.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000066.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 433.0, "y1": 226.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000066" }, { "image": "data/prediction/urban_day_rittenhouse/image/000067.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000067.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 441.0, "y1": 226.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000067" }, { "image": "data/prediction/urban_day_rittenhouse/image/000067.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000067.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.0, "y1": 226.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000067" }, { "image": "data/prediction/urban_day_rittenhouse/image/000068.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000068.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 447.0, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000068" }, { "image": "data/prediction/urban_day_rittenhouse/image/000068.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000068.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 447.0, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000068" }, { "image": "data/prediction/urban_day_rittenhouse/image/000069.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000069.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 448.5, "y1": 236.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000069" }, { "image": "data/prediction/urban_day_rittenhouse/image/000069.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000069.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 448.5, "y1": 236.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000069" }, { "image": "data/prediction/urban_day_rittenhouse/image/000070.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000070.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 467.0, "y1": 242.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000070" }, { "image": "data/prediction/urban_day_rittenhouse/image/000070.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000070.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 467.0, "y1": 242.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000070" }, { "image": "data/prediction/urban_day_rittenhouse/image/000071.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000071.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 478.0, "y1": 236.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000071" }, { "image": "data/prediction/urban_day_rittenhouse/image/000071.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000071.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 478.0, "y1": 236.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000071" }, { "image": "data/prediction/urban_day_rittenhouse/image/000072.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000072.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 536.5, "y1": 206.5, "x2": 639.5, "y2": 314.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000072" }, { "image": "data/prediction/urban_day_rittenhouse/image/000072.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000072.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 536.5, "y1": 206.5, "x2": 639.5, "y2": 314.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000072" }, { "image": "data/prediction/urban_day_rittenhouse/image/000073.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000073.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 533.5, "y1": 207.0, "x2": 639.5, "y2": 316.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000073" }, { "image": "data/prediction/urban_day_rittenhouse/image/000073.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000073.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 533.5, "y1": 207.0, "x2": 639.5, "y2": 316.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000073" }, { "image": "data/prediction/urban_day_rittenhouse/image/000074.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000074.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 535.0, "y1": 207.0, "x2": 639.5, "y2": 317.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000074" }, { "image": "data/prediction/urban_day_rittenhouse/image/000074.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000074.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 535.0, "y1": 207.0, "x2": 639.5, "y2": 317.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000074" }, { "image": "data/prediction/urban_day_rittenhouse/image/000075.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000075.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 536.0, "y1": 210.0, "x2": 639.5, "y2": 322.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000075" }, { "image": "data/prediction/urban_day_rittenhouse/image/000075.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000075.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 536.0, "y1": 210.0, "x2": 639.5, "y2": 322.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000075" }, { "image": "data/prediction/urban_day_rittenhouse/image/000076.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000076.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 545.0, "y1": 210.0, "x2": 639.5, "y2": 324.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000076" }, { "image": "data/prediction/urban_day_rittenhouse/image/000076.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000076.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 545.0, "y1": 210.0, "x2": 639.5, "y2": 324.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000076" }, { "image": "data/prediction/urban_day_rittenhouse/image/000077.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000077.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 546.5, "y1": 217.0, "x2": 639.5, "y2": 334.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000077" }, { "image": "data/prediction/urban_day_rittenhouse/image/000077.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000077.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 546.5, "y1": 217.0, "x2": 639.5, "y2": 334.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000077" }, { "image": "data/prediction/urban_day_rittenhouse/image/000078.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000078.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 564.0, "y1": 214.5, "x2": 639.5, "y2": 337.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000078" }, { "image": "data/prediction/urban_day_rittenhouse/image/000078.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000078.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 564.0, "y1": 214.5, "x2": 639.5, "y2": 337.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000078" }, { "image": "data/prediction/urban_day_rittenhouse/image/000079.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000079.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 577.5, "y1": 219.0, "x2": 639.5, "y2": 352.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000079" }, { "image": "data/prediction/urban_day_rittenhouse/image/000079.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000079.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 577.5, "y1": 219.0, "x2": 639.5, "y2": 352.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000079" }, { "image": "data/prediction/urban_day_rittenhouse/image/000080.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000080.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 586.0, "y1": 221.5, "x2": 639.5, "y2": 359.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000080" }, { "image": "data/prediction/urban_day_rittenhouse/image/000080.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000080.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 586.0, "y1": 221.5, "x2": 639.5, "y2": 359.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000080" }, { "image": "data/prediction/urban_day_rittenhouse/image/000081.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000081.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 604.0, "y1": 224.5, "x2": 639.5, "y2": 394.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000081" }, { "image": "data/prediction/urban_day_rittenhouse/image/000081.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000081.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 604.0, "y1": 224.5, "x2": 639.5, "y2": 394.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000081" }, { "image": "data/prediction/urban_day_rittenhouse/image/000082.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000082.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 27.0, "y1": 179.0, "x2": 126.0, "y2": 216.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000082" }, { "image": "data/prediction/urban_day_rittenhouse/image/000082.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000082.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 27.0, "y1": 179.0, "x2": 126.0, "y2": 216.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000082" }, { "image": "data/prediction/urban_day_rittenhouse/image/000083.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000083.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 28.0, "y1": 179.5, "x2": 128.0, "y2": 216.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000083" }, { "image": "data/prediction/urban_day_rittenhouse/image/000083.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000083.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 28.0, "y1": 179.5, "x2": 128.0, "y2": 216.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000083" }, { "image": "data/prediction/urban_day_rittenhouse/image/000084.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000084.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 25.0, "y1": 181.5, "x2": 125.0, "y2": 218.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000084" }, { "image": "data/prediction/urban_day_rittenhouse/image/000084.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000084.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 25.0, "y1": 181.5, "x2": 125.0, "y2": 218.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000084" }, { "image": "data/prediction/urban_day_rittenhouse/image/000085.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000085.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 25.0, "y1": 183.5, "x2": 125.5, "y2": 220.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000085" }, { "image": "data/prediction/urban_day_rittenhouse/image/000085.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000085.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 25.0, "y1": 183.5, "x2": 125.5, "y2": 220.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000085" }, { "image": "data/prediction/urban_day_rittenhouse/image/000086.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000086.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 20.5, "y1": 184.0, "x2": 123.0, "y2": 221.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000086" }, { "image": "data/prediction/urban_day_rittenhouse/image/000086.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000086.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 20.5, "y1": 184.0, "x2": 123.0, "y2": 221.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000086" }, { "image": "data/prediction/urban_day_rittenhouse/image/000087.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000087.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 13.0, "y1": 178.5, "x2": 117.0, "y2": 217.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000087" }, { "image": "data/prediction/urban_day_rittenhouse/image/000087.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000087.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 13.0, "y1": 178.5, "x2": 117.0, "y2": 217.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000087" }, { "image": "data/prediction/urban_day_rittenhouse/image/000088.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000088.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 8.0, "y1": 179.0, "x2": 115.0, "y2": 219.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000088" }, { "image": "data/prediction/urban_day_rittenhouse/image/000088.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000088.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 8.0, "y1": 179.0, "x2": 115.0, "y2": 219.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000088" }, { "image": "data/prediction/urban_day_rittenhouse/image/000089.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000089.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 12.0, "y1": 185.0, "x2": 118.5, "y2": 225.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000089" }, { "image": "data/prediction/urban_day_rittenhouse/image/000089.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000089.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 12.0, "y1": 185.0, "x2": 118.5, "y2": 225.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000089" }, { "image": "data/prediction/urban_day_rittenhouse/image/000090.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000090.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 389.0, "y1": 206.0, "x2": 528.0, "y2": 313.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000090" }, { "image": "data/prediction/urban_day_rittenhouse/image/000090.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000090.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 389.0, "y1": 206.0, "x2": 528.0, "y2": 313.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000090" }, { "image": "data/prediction/urban_day_rittenhouse/image/000091.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000091.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 389.5, "y1": 205.0, "x2": 529.5, "y2": 313.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000091" }, { "image": "data/prediction/urban_day_rittenhouse/image/000091.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000091.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 389.5, "y1": 205.0, "x2": 529.5, "y2": 313.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000091" }, { "image": "data/prediction/urban_day_rittenhouse/image/000092.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000092.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 388.0, "y1": 205.5, "x2": 527.5, "y2": 315.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000092" }, { "image": "data/prediction/urban_day_rittenhouse/image/000092.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000092.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 388.0, "y1": 205.5, "x2": 527.5, "y2": 315.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000092" }, { "image": "data/prediction/urban_day_rittenhouse/image/000093.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000093.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 390.5, "y1": 209.5, "x2": 529.0, "y2": 320.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000093" }, { "image": "data/prediction/urban_day_rittenhouse/image/000093.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000093.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 390.5, "y1": 209.5, "x2": 529.0, "y2": 320.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000093" }, { "image": "data/prediction/urban_day_rittenhouse/image/000094.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000094.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 393.5, "y1": 210.5, "x2": 540.0, "y2": 326.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000094" }, { "image": "data/prediction/urban_day_rittenhouse/image/000094.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000094.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 393.5, "y1": 210.5, "x2": 540.0, "y2": 326.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000094" }, { "image": "data/prediction/urban_day_rittenhouse/image/000095.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000095.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 397.5, "y1": 210.5, "x2": 552.5, "y2": 332.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000095" }, { "image": "data/prediction/urban_day_rittenhouse/image/000095.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000095.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 397.5, "y1": 210.5, "x2": 552.5, "y2": 332.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000095" }, { "image": "data/prediction/urban_day_rittenhouse/image/000096.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000096.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 404.0, "y1": 210.5, "x2": 573.0, "y2": 342.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000096" }, { "image": "data/prediction/urban_day_rittenhouse/image/000096.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000096.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 404.0, "y1": 210.5, "x2": 573.0, "y2": 342.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000096" }, { "image": "data/prediction/urban_day_rittenhouse/image/000097.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000097.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 414.0, "y1": 213.5, "x2": 595.0, "y2": 357.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000097" }, { "image": "data/prediction/urban_day_rittenhouse/image/000097.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000097.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.0, "y1": 213.5, "x2": 595.0, "y2": 357.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000097" }, { "image": "data/prediction/urban_day_rittenhouse/image/000098.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000098.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.0, "y1": 213.0, "x2": 615.5, "y2": 367.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000098" }, { "image": "data/prediction/urban_day_rittenhouse/image/000098.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000098.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.0, "y1": 213.0, "x2": 615.5, "y2": 367.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000098" }, { "image": "data/prediction/urban_day_rittenhouse/image/000099.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000099.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 424.0, "y1": 217.0, "x2": 636.5, "y2": 387.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000099" }, { "image": "data/prediction/urban_day_rittenhouse/image/000099.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000099.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.0, "y1": 217.0, "x2": 636.5, "y2": 387.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000099" }, { "image": "data/prediction/urban_day_rittenhouse/image/000100.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000100.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 544.5, "y1": 218.0, "x2": 639.5, "y2": 313.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000100" }, { "image": "data/prediction/urban_day_rittenhouse/image/000100.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000100.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 544.5, "y1": 218.0, "x2": 639.5, "y2": 313.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000100" }, { "image": "data/prediction/urban_day_rittenhouse/image/000101.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000101.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 540.5, "y1": 217.5, "x2": 639.5, "y2": 314.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000101" }, { "image": "data/prediction/urban_day_rittenhouse/image/000101.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000101.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 540.5, "y1": 217.5, "x2": 639.5, "y2": 314.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000101" }, { "image": "data/prediction/urban_day_rittenhouse/image/000102.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000102.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 542.0, "y1": 217.5, "x2": 639.5, "y2": 316.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000102" }, { "image": "data/prediction/urban_day_rittenhouse/image/000102.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000102.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 542.0, "y1": 217.5, "x2": 639.5, "y2": 316.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000102" }, { "image": "data/prediction/urban_day_rittenhouse/image/000103.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000103.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 538.5, "y1": 222.5, "x2": 639.5, "y2": 329.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000103" }, { "image": "data/prediction/urban_day_rittenhouse/image/000103.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000103.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 538.5, "y1": 222.5, "x2": 639.5, "y2": 329.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000103" }, { "image": "data/prediction/urban_day_rittenhouse/image/000104.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000104.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 541.5, "y1": 223.0, "x2": 639.5, "y2": 346.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000104" }, { "image": "data/prediction/urban_day_rittenhouse/image/000104.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000104.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 541.5, "y1": 223.0, "x2": 639.5, "y2": 346.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000104" }, { "image": "data/prediction/urban_day_rittenhouse/image/000105.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000105.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 422.0, "y1": 215.5, "x2": 618.5, "y2": 373.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000105" }, { "image": "data/prediction/urban_day_rittenhouse/image/000105.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000105.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 422.0, "y1": 215.5, "x2": 618.5, "y2": 373.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000105" }, { "image": "data/prediction/urban_day_rittenhouse/image/000106.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000106.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 416.0, "y1": 213.0, "x2": 566.0, "y2": 336.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000106" }, { "image": "data/prediction/urban_day_rittenhouse/image/000106.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000106.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.0, "y1": 213.0, "x2": 566.0, "y2": 336.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000106" }, { "image": "data/prediction/urban_day_rittenhouse/image/000107.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000107.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 402.5, "y1": 205.5, "x2": 515.0, "y2": 294.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000107" }, { "image": "data/prediction/urban_day_rittenhouse/image/000107.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000107.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 402.5, "y1": 205.5, "x2": 515.0, "y2": 294.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000107" }, { "image": "data/prediction/urban_day_rittenhouse/image/000108.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000108.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 393.5, "y1": 205.0, "x2": 471.5, "y2": 268.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000108" }, { "image": "data/prediction/urban_day_rittenhouse/image/000108.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000108.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 393.5, "y1": 205.0, "x2": 471.5, "y2": 268.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000108" }, { "image": "data/prediction/urban_day_rittenhouse/image/000109.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000109.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 360.5, "y1": 202.5, "x2": 431.0, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000109" }, { "image": "data/prediction/urban_day_rittenhouse/image/000109.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000109.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 360.5, "y1": 202.5, "x2": 431.0, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000109" }, { "image": "data/prediction/urban_day_rittenhouse/image/000110.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000110.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 346.5, "y1": 201.0, "x2": 401.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000110" }, { "image": "data/prediction/urban_day_rittenhouse/image/000110.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000110.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 346.5, "y1": 201.0, "x2": 401.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000110" }, { "image": "data/prediction/urban_day_rittenhouse/image/000111.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000111.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 185.5, "x2": 108.0, "y2": 229.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000111" }, { "image": "data/prediction/urban_day_rittenhouse/image/000111.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000111.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 185.5, "x2": 108.0, "y2": 229.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000111" }, { "image": "data/prediction/urban_day_rittenhouse/image/000112.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000112.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 189.0, "x2": 105.0, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000112" }, { "image": "data/prediction/urban_day_rittenhouse/image/000112.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000112.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 189.0, "x2": 105.0, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000112" }, { "image": "data/prediction/urban_day_rittenhouse/image/000113.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000113.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 189.5, "x2": 103.0, "y2": 234.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000113" }, { "image": "data/prediction/urban_day_rittenhouse/image/000113.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000113.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 189.5, "x2": 103.0, "y2": 234.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000113" }, { "image": "data/prediction/urban_day_rittenhouse/image/000114.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000114.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 190.5, "x2": 86.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000114" }, { "image": "data/prediction/urban_day_rittenhouse/image/000114.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000114.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 190.5, "x2": 86.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000114" }, { "image": "data/prediction/urban_day_rittenhouse/image/000115.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000115.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 558.5, "y1": 240.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000115" }, { "image": "data/prediction/urban_day_rittenhouse/image/000115.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000115.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 558.5, "y1": 240.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000115" }, { "image": "data/prediction/urban_day_rittenhouse/image/000116.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000116.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 506.5, "y1": 240.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000116" }, { "image": "data/prediction/urban_day_rittenhouse/image/000116.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000116.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 506.5, "y1": 240.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000116" }, { "image": "data/prediction/urban_day_rittenhouse/image/000117.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000117.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 489.5, "y1": 232.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000117" }, { "image": "data/prediction/urban_day_rittenhouse/image/000117.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000117.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 489.5, "y1": 232.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000117" }, { "image": "data/prediction/urban_day_rittenhouse/image/000118.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000118.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 466.0, "y1": 226.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000118" }, { "image": "data/prediction/urban_day_rittenhouse/image/000118.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000118.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 466.0, "y1": 226.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000118" }, { "image": "data/prediction/urban_day_rittenhouse/image/000119.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000119.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 417.0, "y1": 220.0, "x2": 625.0, "y2": 376.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000119" }, { "image": "data/prediction/urban_day_rittenhouse/image/000119.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000119.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 417.0, "y1": 220.0, "x2": 625.0, "y2": 376.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000119" }, { "image": "data/prediction/urban_day_rittenhouse/image/000120.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000120.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 384.5, "y1": 211.5, "x2": 559.5, "y2": 341.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000120" }, { "image": "data/prediction/urban_day_rittenhouse/image/000120.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000120.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 384.5, "y1": 211.5, "x2": 559.5, "y2": 341.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000120" }, { "image": "data/prediction/urban_day_rittenhouse/image/000121.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000121.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 364.5, "y1": 209.5, "x2": 510.5, "y2": 319.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000121" }, { "image": "data/prediction/urban_day_rittenhouse/image/000121.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000121.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 364.5, "y1": 209.5, "x2": 510.5, "y2": 319.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000121" }, { "image": "data/prediction/urban_day_rittenhouse/image/000122.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000122.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 351.5, "y1": 211.0, "x2": 463.0, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000122" }, { "image": "data/prediction/urban_day_rittenhouse/image/000122.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000122.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 351.5, "y1": 211.0, "x2": 463.0, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000122" }, { "image": "data/prediction/urban_day_rittenhouse/image/000123.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000123.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 124.5, "x2": 64.5, "y2": 286.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000123" }, { "image": "data/prediction/urban_day_rittenhouse/image/000123.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000123.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 124.5, "x2": 64.5, "y2": 286.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000123" }, { "image": "data/prediction/urban_day_rittenhouse/image/000124.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000124.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 85.5, "x2": 143.5, "y2": 351.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000124" }, { "image": "data/prediction/urban_day_rittenhouse/image/000124.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000124.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 85.5, "x2": 143.5, "y2": 351.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000124" }, { "image": "data/prediction/urban_day_rittenhouse/image/000125.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000125.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 10.0, "y1": 204.0, "x2": 90.0, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000125" }, { "image": "data/prediction/urban_day_rittenhouse/image/000125.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000125.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 10.0, "y1": 204.0, "x2": 90.0, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000125" }, { "image": "data/prediction/urban_day_rittenhouse/image/000126.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000126.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 97.5, "y1": 209.5, "x2": 204.5, "y2": 286.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000126" }, { "image": "data/prediction/urban_day_rittenhouse/image/000126.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000126.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 97.5, "y1": 209.5, "x2": 204.5, "y2": 286.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000126" }, { "image": "data/prediction/urban_day_rittenhouse/image/000127.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000127.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 138.0, "y1": 208.5, "x2": 252.0, "y2": 293.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000127" }, { "image": "data/prediction/urban_day_rittenhouse/image/000127.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000127.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 138.0, "y1": 208.5, "x2": 252.0, "y2": 293.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000127" }, { "image": "data/prediction/urban_day_rittenhouse/image/000128.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000128.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 220.5, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000128" }, { "image": "data/prediction/urban_day_rittenhouse/image/000128.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000128.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 220.5, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000128" }, { "image": "data/prediction/urban_day_rittenhouse/image/000129.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000129.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 22.5, "y1": 217.0, "x2": 231.0, "y2": 368.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000129" }, { "image": "data/prediction/urban_day_rittenhouse/image/000129.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000129.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 22.5, "y1": 217.0, "x2": 231.0, "y2": 368.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000129" }, { "image": "data/prediction/urban_day_rittenhouse/image/000130.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000130.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 197.5, "y1": 146.5, "x2": 292.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000130" }, { "image": "data/prediction/urban_day_rittenhouse/image/000130.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000130.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 197.5, "y1": 146.5, "x2": 292.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000130" }, { "image": "data/prediction/urban_day_rittenhouse/image/000131.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000131.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 117.0, "y1": 123.0, "x2": 281.5, "y2": 291.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000131" }, { "image": "data/prediction/urban_day_rittenhouse/image/000131.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000131.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 117.0, "y1": 123.0, "x2": 281.5, "y2": 291.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000131" }, { "image": "data/prediction/urban_day_rittenhouse/image/000132.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000132.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 26.5, "x2": 233.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000132" }, { "image": "data/prediction/urban_day_rittenhouse/image/000132.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000132.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 26.5, "x2": 233.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000132" }, { "image": "data/prediction/urban_day_rittenhouse/image/000133.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000133.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 230.5, "y1": 193.5, "x2": 291.5, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000133" }, { "image": "data/prediction/urban_day_rittenhouse/image/000133.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000133.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 230.5, "y1": 193.5, "x2": 291.5, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000133" }, { "image": "data/prediction/urban_day_rittenhouse/image/000134.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000134.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 184.5, "y1": 198.5, "x2": 272.5, "y2": 265.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000134" }, { "image": "data/prediction/urban_day_rittenhouse/image/000134.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000134.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 184.5, "y1": 198.5, "x2": 272.5, "y2": 265.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000134" }, { "image": "data/prediction/urban_day_rittenhouse/image/000135.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000135.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 71.5, "y1": 200.0, "x2": 240.0, "y2": 326.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000135" }, { "image": "data/prediction/urban_day_rittenhouse/image/000135.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000135.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 71.5, "y1": 200.0, "x2": 240.0, "y2": 326.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000135" }, { "image": "data/prediction/urban_day_rittenhouse/image/000136.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000136.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 245.5, "y1": 182.5, "x2": 303.5, "y2": 231.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000136" }, { "image": "data/prediction/urban_day_rittenhouse/image/000136.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000136.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 245.5, "y1": 182.5, "x2": 303.5, "y2": 231.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000136" }, { "image": "data/prediction/urban_day_rittenhouse/image/000137.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000137.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 215.0, "y1": 186.5, "x2": 282.5, "y2": 246.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000137" }, { "image": "data/prediction/urban_day_rittenhouse/image/000137.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000137.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 215.0, "y1": 186.5, "x2": 282.5, "y2": 246.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000137" }, { "image": "data/prediction/urban_day_rittenhouse/image/000138.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000138.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 146.5, "y1": 179.5, "x2": 268.5, "y2": 282.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000138" }, { "image": "data/prediction/urban_day_rittenhouse/image/000138.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000138.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 146.5, "y1": 179.5, "x2": 268.5, "y2": 282.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000138" }, { "image": "data/prediction/urban_day_rittenhouse/image/000139.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000139.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 230.0, "y1": 189.5, "x2": 288.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000139" }, { "image": "data/prediction/urban_day_rittenhouse/image/000139.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000139.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 230.0, "y1": 189.5, "x2": 288.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000139" }, { "image": "data/prediction/urban_day_rittenhouse/image/000140.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000140.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 188.0, "y1": 187.0, "x2": 279.5, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000140" }, { "image": "data/prediction/urban_day_rittenhouse/image/000140.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000140.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 188.0, "y1": 187.0, "x2": 279.5, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000140" }, { "image": "data/prediction/urban_day_rittenhouse/image/000141.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000141.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 59.5, "y1": 184.0, "x2": 250.5, "y2": 325.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000141" }, { "image": "data/prediction/urban_day_rittenhouse/image/000141.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000141.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 59.5, "y1": 184.0, "x2": 250.5, "y2": 325.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000141" }, { "image": "data/prediction/urban_day_rittenhouse/image/000142.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000142.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 227.5, "y1": 197.0, "x2": 288.5, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000142" }, { "image": "data/prediction/urban_day_rittenhouse/image/000142.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000142.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 227.5, "y1": 197.0, "x2": 288.5, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000142" }, { "image": "data/prediction/urban_day_rittenhouse/image/000143.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000143.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 166.0, "y1": 201.0, "x2": 267.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000143" }, { "image": "data/prediction/urban_day_rittenhouse/image/000143.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000143.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 166.0, "y1": 201.0, "x2": 267.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000143" }, { "image": "data/prediction/urban_day_rittenhouse/image/000144.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000144.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 12.0, "y1": 207.5, "x2": 234.5, "y2": 367.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000144" }, { "image": "data/prediction/urban_day_rittenhouse/image/000144.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000144.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 12.0, "y1": 207.5, "x2": 234.5, "y2": 367.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000144" }, { "image": "data/prediction/urban_day_rittenhouse/image/000145.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000145.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 195.0, "y1": 202.0, "x2": 273.5, "y2": 261.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000145" }, { "image": "data/prediction/urban_day_rittenhouse/image/000145.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000145.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 195.0, "y1": 202.0, "x2": 273.5, "y2": 261.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000145" }, { "image": "data/prediction/urban_day_rittenhouse/image/000146.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000146.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 80.0, "y1": 208.5, "x2": 230.5, "y2": 321.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000146" }, { "image": "data/prediction/urban_day_rittenhouse/image/000146.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000146.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 80.0, "y1": 208.5, "x2": 230.5, "y2": 321.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000146" }, { "image": "data/prediction/urban_day_rittenhouse/image/000147.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000147.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 207.5, "y1": 187.0, "x2": 279.5, "y2": 243.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000147" }, { "image": "data/prediction/urban_day_rittenhouse/image/000147.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000147.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 207.5, "y1": 187.0, "x2": 279.5, "y2": 243.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000147" }, { "image": "data/prediction/urban_day_rittenhouse/image/000148.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000148.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 122.0, "y1": 186.5, "x2": 247.5, "y2": 284.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000148" }, { "image": "data/prediction/urban_day_rittenhouse/image/000148.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000148.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 122.0, "y1": 186.5, "x2": 247.5, "y2": 284.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000148" }, { "image": "data/prediction/urban_day_rittenhouse/image/000149.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000149.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 210.5, "y1": 193.5, "x2": 277.0, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000149" }, { "image": "data/prediction/urban_day_rittenhouse/image/000149.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000149.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 210.5, "y1": 193.5, "x2": 277.0, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000149" }, { "image": "data/prediction/urban_day_rittenhouse/image/000150.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000150.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 150.5, "y1": 198.0, "x2": 259.0, "y2": 279.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000150" }, { "image": "data/prediction/urban_day_rittenhouse/image/000150.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000150.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 150.5, "y1": 198.0, "x2": 259.0, "y2": 279.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000150" }, { "image": "data/prediction/urban_day_rittenhouse/image/000151.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000151.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 247.5, "y1": 199.0, "x2": 299.5, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000151" }, { "image": "data/prediction/urban_day_rittenhouse/image/000151.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000151.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 247.5, "y1": 199.0, "x2": 299.5, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000151" }, { "image": "data/prediction/urban_day_rittenhouse/image/000152.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000152.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 201.0, "y1": 196.5, "x2": 279.0, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000152" }, { "image": "data/prediction/urban_day_rittenhouse/image/000152.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000152.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 201.0, "y1": 196.5, "x2": 279.0, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000152" }, { "image": "data/prediction/urban_day_rittenhouse/image/000153.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000153.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 121.0, "y1": 200.5, "x2": 260.0, "y2": 302.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000153" }, { "image": "data/prediction/urban_day_rittenhouse/image/000153.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000153.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 121.0, "y1": 200.5, "x2": 260.0, "y2": 302.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000153" }, { "image": "data/prediction/urban_day_rittenhouse/image/000154.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000154.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 244.5, "y1": 192.5, "x2": 296.0, "y2": 231.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000154" }, { "image": "data/prediction/urban_day_rittenhouse/image/000154.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000154.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 244.5, "y1": 192.5, "x2": 296.0, "y2": 231.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000154" }, { "image": "data/prediction/urban_day_rittenhouse/image/000155.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000155.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 213.5, "y1": 195.5, "x2": 279.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000155" }, { "image": "data/prediction/urban_day_rittenhouse/image/000155.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000155.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 213.5, "y1": 195.5, "x2": 279.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000155" }, { "image": "data/prediction/urban_day_rittenhouse/image/000156.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000156.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 237.5, "y1": 196.0, "x2": 291.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000156" }, { "image": "data/prediction/urban_day_rittenhouse/image/000156.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000156.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 237.5, "y1": 196.0, "x2": 291.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000156" }, { "image": "data/prediction/urban_day_rittenhouse/image/000157.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000157.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 51.5, "y1": 203.5, "x2": 244.0, "y2": 346.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000157" }, { "image": "data/prediction/urban_day_rittenhouse/image/000157.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000157.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 51.5, "y1": 203.5, "x2": 244.0, "y2": 346.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000157" }, { "image": "data/prediction/urban_day_rittenhouse/image/000158.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000158.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 361.0, "y1": 154.5, "x2": 405.0, "y2": 205.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000158" }, { "image": "data/prediction/urban_day_rittenhouse/image/000158.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000158.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 361.0, "y1": 154.5, "x2": 405.0, "y2": 205.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000158" }, { "image": "data/prediction/urban_day_rittenhouse/image/000159.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000159.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 361.5, "y1": 155.5, "x2": 411.0, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000159" }, { "image": "data/prediction/urban_day_rittenhouse/image/000159.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000159.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 361.5, "y1": 155.5, "x2": 411.0, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000159" }, { "image": "data/prediction/urban_day_rittenhouse/image/000160.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000160.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 362.5, "y1": 154.5, "x2": 414.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000160" }, { "image": "data/prediction/urban_day_rittenhouse/image/000160.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000160.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 154.5, "x2": 414.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000160" }, { "image": "data/prediction/urban_day_rittenhouse/image/000161.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000161.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.5, "y1": 154.5, "x2": 416.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000161" }, { "image": "data/prediction/urban_day_rittenhouse/image/000161.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000161.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 154.5, "x2": 416.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000161" }, { "image": "data/prediction/urban_day_rittenhouse/image/000162.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000162.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000162" }, { "image": "data/prediction/urban_day_rittenhouse/image/000162.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000162.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000162" }, { "image": "data/prediction/urban_day_rittenhouse/image/000163.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000163.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 415.5, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000163" }, { "image": "data/prediction/urban_day_rittenhouse/image/000163.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000163.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 415.5, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000163" }, { "image": "data/prediction/urban_day_rittenhouse/image/000164.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000164.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000164" }, { "image": "data/prediction/urban_day_rittenhouse/image/000164.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000164.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000164" }, { "image": "data/prediction/urban_day_rittenhouse/image/000165.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000165.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 363.0, "y1": 155.0, "x2": 415.5, "y2": 214.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000165" }, { "image": "data/prediction/urban_day_rittenhouse/image/000165.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000165.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 363.0, "y1": 155.0, "x2": 415.5, "y2": 214.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000165" }, { "image": "data/prediction/urban_day_rittenhouse/image/000166.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000166.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 164.5, "y1": 202.0, "x2": 272.0, "y2": 283.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000166" }, { "image": "data/prediction/urban_day_rittenhouse/image/000166.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000166.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 164.5, "y1": 202.0, "x2": 272.0, "y2": 283.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000166" }, { "image": "data/prediction/urban_day_rittenhouse/image/000167.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000167.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 29.0, "y1": 212.0, "x2": 243.5, "y2": 367.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000167" }, { "image": "data/prediction/urban_day_rittenhouse/image/000167.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000167.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 29.0, "y1": 212.0, "x2": 243.5, "y2": 367.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000167" }, { "image": "data/prediction/urban_day_rittenhouse/image/000168.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000168.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 308.0, "y1": 188.0, "x2": 358.5, "y2": 234.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000168" }, { "image": "data/prediction/urban_day_rittenhouse/image/000168.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000168.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 308.0, "y1": 188.0, "x2": 358.5, "y2": 234.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000168" }, { "image": "data/prediction/urban_day_rittenhouse/image/000169.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000169.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 305.5, "y1": 189.5, "x2": 358.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000169" }, { "image": "data/prediction/urban_day_rittenhouse/image/000169.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000169.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.5, "y1": 189.5, "x2": 358.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000169" }, { "image": "data/prediction/urban_day_rittenhouse/image/000170.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000170.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 307.5, "y1": 190.0, "x2": 362.0, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000170" }, { "image": "data/prediction/urban_day_rittenhouse/image/000170.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000170.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.5, "y1": 190.0, "x2": 362.0, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000170" }, { "image": "data/prediction/urban_day_rittenhouse/image/000171.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000171.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 306.0, "y1": 191.0, "x2": 361.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000171" }, { "image": "data/prediction/urban_day_rittenhouse/image/000171.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000171.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.0, "y1": 191.0, "x2": 361.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000171" }, { "image": "data/prediction/urban_day_rittenhouse/image/000172.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000172.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 306.5, "y1": 190.5, "x2": 362.0, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000172" }, { "image": "data/prediction/urban_day_rittenhouse/image/000172.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000172.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.5, "y1": 190.5, "x2": 362.0, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000172" }, { "image": "data/prediction/urban_day_rittenhouse/image/000173.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000173.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 307.5, "y1": 191.0, "x2": 363.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000173" }, { "image": "data/prediction/urban_day_rittenhouse/image/000173.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000173.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.5, "y1": 191.0, "x2": 363.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000173" }, { "image": "data/prediction/urban_day_rittenhouse/image/000174.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000174.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 307.5, "y1": 190.5, "x2": 363.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000174" }, { "image": "data/prediction/urban_day_rittenhouse/image/000174.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000174.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.5, "y1": 190.5, "x2": 363.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000174" }, { "image": "data/prediction/urban_day_rittenhouse/image/000175.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000175.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 324.0, "y1": 180.0, "x2": 379.0, "y2": 207.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000175" }, { "image": "data/prediction/urban_day_rittenhouse/image/000175.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000175.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.0, "y1": 180.0, "x2": 379.0, "y2": 207.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000175" }, { "image": "data/prediction/urban_day_rittenhouse/image/000176.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000176.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 306.0, "y1": 190.5, "x2": 361.5, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000176" }, { "image": "data/prediction/urban_day_rittenhouse/image/000176.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000176.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.0, "y1": 190.5, "x2": 361.5, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000176" }, { "image": "data/prediction/urban_day_rittenhouse/image/000177.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000177.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 307.0, "y1": 190.5, "x2": 360.5, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000177" }, { "image": "data/prediction/urban_day_rittenhouse/image/000177.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000177.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.0, "y1": 190.5, "x2": 360.5, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000177" }, { "image": "data/prediction/urban_day_rittenhouse/image/000178.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000178.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 361.0, "y1": 154.5, "x2": 418.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000178" }, { "image": "data/prediction/urban_day_rittenhouse/image/000178.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000178.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 361.0, "y1": 154.5, "x2": 418.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000178" }, { "image": "data/prediction/urban_day_rittenhouse/image/000179.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000179.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 363.0, "y1": 154.0, "x2": 416.5, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000179" }, { "image": "data/prediction/urban_day_rittenhouse/image/000179.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000179.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 363.0, "y1": 154.0, "x2": 416.5, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000179" }, { "image": "data/prediction/urban_day_rittenhouse/image/000180.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000180.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000180" }, { "image": "data/prediction/urban_day_rittenhouse/image/000180.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000180.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000180" }, { "image": "data/prediction/urban_day_rittenhouse/image/000181.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000181.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 363.0, "y1": 155.0, "x2": 416.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000181" }, { "image": "data/prediction/urban_day_rittenhouse/image/000181.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000181.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 363.0, "y1": 155.0, "x2": 416.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000181" }, { "image": "data/prediction/urban_day_rittenhouse/image/000182.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000182.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 590.0, "y1": 190.5, "x2": 639.5, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000182" }, { "image": "data/prediction/urban_day_rittenhouse/image/000182.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000182.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 590.0, "y1": 190.5, "x2": 639.5, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000182" }, { "image": "data/prediction/urban_day_rittenhouse/image/000183.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000183.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 514.5, "y1": 176.5, "x2": 639.5, "y2": 256.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000183" }, { "image": "data/prediction/urban_day_rittenhouse/image/000183.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000183.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 514.5, "y1": 176.5, "x2": 639.5, "y2": 256.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000183" }, { "image": "data/prediction/urban_day_rittenhouse/image/000184.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000184.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 438.0, "y1": 174.0, "x2": 633.5, "y2": 257.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000184" }, { "image": "data/prediction/urban_day_rittenhouse/image/000184.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000184.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 438.0, "y1": 174.0, "x2": 633.5, "y2": 257.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000184" }, { "image": "data/prediction/urban_day_rittenhouse/image/000185.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000185.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 373.0, "y1": 174.0, "x2": 565.0, "y2": 254.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000185" }, { "image": "data/prediction/urban_day_rittenhouse/image/000185.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000185.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 373.0, "y1": 174.0, "x2": 565.0, "y2": 254.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000185" }, { "image": "data/prediction/urban_day_rittenhouse/image/000186.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000186.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 336.0, "y1": 174.5, "x2": 504.0, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000186" }, { "image": "data/prediction/urban_day_rittenhouse/image/000186.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000186.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 336.0, "y1": 174.5, "x2": 504.0, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000186" }, { "image": "data/prediction/urban_day_rittenhouse/image/000187.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000187.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 316.5, "y1": 176.0, "x2": 454.0, "y2": 244.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000187" }, { "image": "data/prediction/urban_day_rittenhouse/image/000187.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000187.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 316.5, "y1": 176.0, "x2": 454.0, "y2": 244.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000187" }, { "image": "data/prediction/urban_day_rittenhouse/image/000188.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000188.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 309.5, "y1": 174.5, "x2": 416.0, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000188" }, { "image": "data/prediction/urban_day_rittenhouse/image/000188.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000188.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.5, "y1": 174.5, "x2": 416.0, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000188" }, { "image": "data/prediction/urban_day_rittenhouse/image/000189.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000189.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 306.5, "y1": 174.0, "x2": 390.0, "y2": 228.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000189" }, { "image": "data/prediction/urban_day_rittenhouse/image/000189.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000189.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.5, "y1": 174.0, "x2": 390.0, "y2": 228.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000189" }, { "image": "data/prediction/urban_day_rittenhouse/image/000190.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000190.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 304.5, "y1": 176.0, "x2": 366.5, "y2": 221.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000190" }, { "image": "data/prediction/urban_day_rittenhouse/image/000190.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000190.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.5, "y1": 176.0, "x2": 366.5, "y2": 221.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000190" }, { "image": "data/prediction/urban_day_rittenhouse/image/000191.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000191.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 574.5, "y1": 188.5, "x2": 639.5, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000191" }, { "image": "data/prediction/urban_day_rittenhouse/image/000191.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000191.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 574.5, "y1": 188.5, "x2": 639.5, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000191" }, { "image": "data/prediction/urban_day_rittenhouse/image/000192.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000192.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 477.0, "y1": 188.5, "x2": 639.5, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000192" }, { "image": "data/prediction/urban_day_rittenhouse/image/000192.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000192.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 477.0, "y1": 188.5, "x2": 639.5, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000192" }, { "image": "data/prediction/urban_day_rittenhouse/image/000193.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000193.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 373.0, "y1": 186.0, "x2": 578.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000193" }, { "image": "data/prediction/urban_day_rittenhouse/image/000193.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000193.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 373.0, "y1": 186.0, "x2": 578.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000193" }, { "image": "data/prediction/urban_day_rittenhouse/image/000194.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000194.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 318.0, "y1": 187.0, "x2": 504.5, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000194" }, { "image": "data/prediction/urban_day_rittenhouse/image/000194.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000194.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 318.0, "y1": 187.0, "x2": 504.5, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000194" }, { "image": "data/prediction/urban_day_rittenhouse/image/000195.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000195.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 301.0, "y1": 186.0, "x2": 454.0, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000195" }, { "image": "data/prediction/urban_day_rittenhouse/image/000195.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000195.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 301.0, "y1": 186.0, "x2": 454.0, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000195" }, { "image": "data/prediction/urban_day_rittenhouse/image/000196.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000196.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 302.5, "y1": 184.5, "x2": 412.5, "y2": 244.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000196" }, { "image": "data/prediction/urban_day_rittenhouse/image/000196.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000196.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 302.5, "y1": 184.5, "x2": 412.5, "y2": 244.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000196" }, { "image": "data/prediction/urban_day_rittenhouse/image/000197.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000197.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 303.0, "y1": 184.5, "x2": 385.0, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000197" }, { "image": "data/prediction/urban_day_rittenhouse/image/000197.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000197.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 303.0, "y1": 184.5, "x2": 385.0, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000197" }, { "image": "data/prediction/urban_day_rittenhouse/image/000198.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000198.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 299.0, "y1": 185.0, "x2": 365.0, "y2": 231.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000198" }, { "image": "data/prediction/urban_day_rittenhouse/image/000198.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000198.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.0, "y1": 185.0, "x2": 365.0, "y2": 231.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000198" }, { "image": "data/prediction/urban_day_rittenhouse/image/000199.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000199.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 364.0, "y1": 154.5, "x2": 414.0, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000199" }, { "image": "data/prediction/urban_day_rittenhouse/image/000199.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000199.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 364.0, "y1": 154.5, "x2": 414.0, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000199" }, { "image": "data/prediction/urban_day_rittenhouse/image/000200.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000200.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 364.5, "y1": 155.5, "x2": 414.0, "y2": 211.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000200" }, { "image": "data/prediction/urban_day_rittenhouse/image/000200.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000200.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 364.5, "y1": 155.5, "x2": 414.0, "y2": 211.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000200" }, { "image": "data/prediction/urban_day_rittenhouse/image/000201.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000201.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 415.5, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000201" }, { "image": "data/prediction/urban_day_rittenhouse/image/000201.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000201.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 155.0, "x2": 415.5, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000201" }, { "image": "data/prediction/urban_day_rittenhouse/image/000202.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000202.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.0, "y1": 154.5, "x2": 414.5, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000202" }, { "image": "data/prediction/urban_day_rittenhouse/image/000202.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000202.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.0, "y1": 154.5, "x2": 414.5, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000202" }, { "image": "data/prediction/urban_day_rittenhouse/image/000203.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000203.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 362.0, "y1": 154.5, "x2": 416.5, "y2": 214.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000203" }, { "image": "data/prediction/urban_day_rittenhouse/image/000203.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000203.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.0, "y1": 154.5, "x2": 416.5, "y2": 214.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000203" }, { "image": "data/prediction/urban_day_rittenhouse/image/000204.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000204.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 440.5, "y1": 137.0, "x2": 639.5, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000204" }, { "image": "data/prediction/urban_day_rittenhouse/image/000204.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000204.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 440.5, "y1": 137.0, "x2": 639.5, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000204" }, { "image": "data/prediction/urban_day_rittenhouse/image/000205.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000205.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 353.0, "y1": 136.5, "x2": 639.5, "y2": 274.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000205" }, { "image": "data/prediction/urban_day_rittenhouse/image/000205.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000205.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 353.0, "y1": 136.5, "x2": 639.5, "y2": 274.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000205" }, { "image": "data/prediction/urban_day_rittenhouse/image/000206.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000206.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 278.5, "y1": 135.0, "x2": 639.5, "y2": 268.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000206" }, { "image": "data/prediction/urban_day_rittenhouse/image/000206.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000206.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 278.5, "y1": 135.0, "x2": 639.5, "y2": 268.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000206" }, { "image": "data/prediction/urban_day_rittenhouse/image/000207.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000207.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 247.0, "y1": 135.0, "x2": 639.5, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000207" }, { "image": "data/prediction/urban_day_rittenhouse/image/000207.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000207.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 247.0, "y1": 135.0, "x2": 639.5, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000207" }, { "image": "data/prediction/urban_day_rittenhouse/image/000208.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000208.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 241.5, "y1": 131.5, "x2": 639.5, "y2": 273.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000208" }, { "image": "data/prediction/urban_day_rittenhouse/image/000208.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000208.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 241.5, "y1": 131.5, "x2": 639.5, "y2": 273.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000208" }, { "image": "data/prediction/urban_day_rittenhouse/image/000209.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000209.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 252.0, "y1": 129.0, "x2": 607.0, "y2": 272.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000209" }, { "image": "data/prediction/urban_day_rittenhouse/image/000209.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000209.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 252.0, "y1": 129.0, "x2": 607.0, "y2": 272.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000209" }, { "image": "data/prediction/urban_day_rittenhouse/image/000210.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000210.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 337.5, "y1": 132.0, "x2": 639.5, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000210" }, { "image": "data/prediction/urban_day_rittenhouse/image/000210.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000210.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 337.5, "y1": 132.0, "x2": 639.5, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000210" }, { "image": "data/prediction/urban_day_rittenhouse/image/000211.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000211.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 289.0, "y1": 133.5, "x2": 541.5, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000211" }, { "image": "data/prediction/urban_day_rittenhouse/image/000211.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000211.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 289.0, "y1": 133.5, "x2": 541.5, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000211" }, { "image": "data/prediction/urban_day_rittenhouse/image/000212.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000212.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 287.0, "y1": 145.0, "x2": 443.0, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000212" }, { "image": "data/prediction/urban_day_rittenhouse/image/000212.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000212.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 287.0, "y1": 145.0, "x2": 443.0, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000212" }, { "image": "data/prediction/urban_day_rittenhouse/image/000213.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000213.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 297.5, "y1": 152.5, "x2": 384.0, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000213" }, { "image": "data/prediction/urban_day_rittenhouse/image/000213.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000213.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 297.5, "y1": 152.5, "x2": 384.0, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000213" }, { "image": "data/prediction/urban_day_rittenhouse/image/000214.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000214.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 304.5, "y1": 159.5, "x2": 364.0, "y2": 228.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000214" }, { "image": "data/prediction/urban_day_rittenhouse/image/000214.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000214.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.5, "y1": 159.5, "x2": 364.0, "y2": 228.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000214" }, { "image": "data/prediction/urban_day_rittenhouse/image/000215.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000215.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 302.5, "y1": 160.5, "x2": 354.0, "y2": 220.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000215" }, { "image": "data/prediction/urban_day_rittenhouse/image/000215.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000215.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 302.5, "y1": 160.5, "x2": 354.0, "y2": 220.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000215" }, { "image": "data/prediction/urban_day_rittenhouse/image/000216.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000216.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 370.0, "y1": 156.5, "x2": 431.0, "y2": 214.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000216" }, { "image": "data/prediction/urban_day_rittenhouse/image/000216.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000216.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 370.0, "y1": 156.5, "x2": 431.0, "y2": 214.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000216" }, { "image": "data/prediction/urban_day_rittenhouse/image/000217.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000217.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 360.0, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000217" }, { "image": "data/prediction/urban_day_rittenhouse/image/000217.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000217.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 360.0, "y1": 155.0, "x2": 416.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000217" }, { "image": "data/prediction/urban_day_rittenhouse/image/000218.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000218.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 363.0, "y1": 154.5, "x2": 416.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000218" }, { "image": "data/prediction/urban_day_rittenhouse/image/000218.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000218.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 363.0, "y1": 154.5, "x2": 416.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000218" }, { "image": "data/prediction/urban_day_rittenhouse/image/000219.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000219.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 362.5, "y1": 154.5, "x2": 413.5, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000219" }, { "image": "data/prediction/urban_day_rittenhouse/image/000219.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000219.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.5, "y1": 154.5, "x2": 413.5, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000219" }, { "image": "data/prediction/urban_day_rittenhouse/image/000220.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000220.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 362.0, "y1": 156.0, "x2": 417.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000220" }, { "image": "data/prediction/urban_day_rittenhouse/image/000220.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000220.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.0, "y1": 156.0, "x2": 417.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000220" }, { "image": "data/prediction/urban_day_rittenhouse/image/000221.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000221.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 367.5, "y1": 161.5, "x2": 421.0, "y2": 220.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000221" }, { "image": "data/prediction/urban_day_rittenhouse/image/000221.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000221.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 367.5, "y1": 161.5, "x2": 421.0, "y2": 220.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000221" }, { "image": "data/prediction/urban_day_rittenhouse/image/000222.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000222.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 387.5, "y1": 163.5, "x2": 443.0, "y2": 222.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000222" }, { "image": "data/prediction/urban_day_rittenhouse/image/000222.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000222.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 387.5, "y1": 163.5, "x2": 443.0, "y2": 222.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000222" }, { "image": "data/prediction/urban_day_rittenhouse/image/000223.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000223.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 546.0, "y1": 199.0, "x2": 639.5, "y2": 264.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000223" }, { "image": "data/prediction/urban_day_rittenhouse/image/000223.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000223.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 546.0, "y1": 199.0, "x2": 639.5, "y2": 264.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000223" }, { "image": "data/prediction/urban_day_rittenhouse/image/000224.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000224.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 423.0, "y1": 197.5, "x2": 589.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000224" }, { "image": "data/prediction/urban_day_rittenhouse/image/000224.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000224.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 423.0, "y1": 197.5, "x2": 589.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000224" }, { "image": "data/prediction/urban_day_rittenhouse/image/000225.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000225.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 352.0, "y1": 198.5, "x2": 497.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000225" }, { "image": "data/prediction/urban_day_rittenhouse/image/000225.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000225.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 352.0, "y1": 198.5, "x2": 497.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000225" }, { "image": "data/prediction/urban_day_rittenhouse/image/000226.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000226.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 318.0, "y1": 195.0, "x2": 412.0, "y2": 244.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000226" }, { "image": "data/prediction/urban_day_rittenhouse/image/000226.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000226.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 318.0, "y1": 195.0, "x2": 412.0, "y2": 244.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000226" }, { "image": "data/prediction/urban_day_rittenhouse/image/000227.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000227.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 309.0, "y1": 195.0, "x2": 368.5, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000227" }, { "image": "data/prediction/urban_day_rittenhouse/image/000227.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000227.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.0, "y1": 195.0, "x2": 368.5, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000227" }, { "image": "data/prediction/urban_day_rittenhouse/image/000228.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000228.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 291.5, "y1": 168.5, "x2": 349.5, "y2": 216.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000228" }, { "image": "data/prediction/urban_day_rittenhouse/image/000228.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000228.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 291.5, "y1": 168.5, "x2": 349.5, "y2": 216.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000228" }, { "image": "data/prediction/urban_day_rittenhouse/image/000229.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000229.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 300.5, "y1": 170.0, "x2": 368.5, "y2": 216.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000229" }, { "image": "data/prediction/urban_day_rittenhouse/image/000229.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000229.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 300.5, "y1": 170.0, "x2": 368.5, "y2": 216.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000229" }, { "image": "data/prediction/urban_day_rittenhouse/image/000230.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000230.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 416.5, "y1": 159.5, "x2": 481.0, "y2": 226.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000230" }, { "image": "data/prediction/urban_day_rittenhouse/image/000230.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000230.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.5, "y1": 159.5, "x2": 481.0, "y2": 226.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000230" }, { "image": "data/prediction/urban_day_rittenhouse/image/000231.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000231.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 416.5, "y1": 156.5, "x2": 493.5, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000231" }, { "image": "data/prediction/urban_day_rittenhouse/image/000231.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000231.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.5, "y1": 156.5, "x2": 493.5, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000231" }, { "image": "data/prediction/urban_day_rittenhouse/image/000232.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000232.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 413.0, "y1": 152.5, "x2": 515.0, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000232" }, { "image": "data/prediction/urban_day_rittenhouse/image/000232.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000232.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 413.0, "y1": 152.5, "x2": 515.0, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000232" }, { "image": "data/prediction/urban_day_rittenhouse/image/000233.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000233.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 421.0, "y1": 124.5, "x2": 574.5, "y2": 269.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000233" }, { "image": "data/prediction/urban_day_rittenhouse/image/000233.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000233.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.0, "y1": 124.5, "x2": 574.5, "y2": 269.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000233" }, { "image": "data/prediction/urban_day_rittenhouse/image/000234.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000234.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 449.5, "y1": 68.0, "x2": 639.5, "y2": 336.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000234" }, { "image": "data/prediction/urban_day_rittenhouse/image/000234.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000234.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 449.5, "y1": 68.0, "x2": 639.5, "y2": 336.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000234" }, { "image": "data/prediction/urban_day_rittenhouse/image/000235.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000235.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 3.5, "y1": 160.5, "x2": 193.0, "y2": 222.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000235" }, { "image": "data/prediction/urban_day_rittenhouse/image/000235.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000235.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 3.5, "y1": 160.5, "x2": 193.0, "y2": 222.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000235" }, { "image": "data/prediction/urban_day_rittenhouse/image/000236.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000236.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 117.5, "y1": 185.5, "x2": 197.0, "y2": 216.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000236" }, { "image": "data/prediction/urban_day_rittenhouse/image/000236.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000236.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 117.5, "y1": 185.5, "x2": 197.0, "y2": 216.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000236" }, { "image": "data/prediction/urban_day_rittenhouse/image/000237.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000237.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 397.0, "y1": 191.5, "x2": 467.5, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000237" }, { "image": "data/prediction/urban_day_rittenhouse/image/000237.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000237.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 397.0, "y1": 191.5, "x2": 467.5, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000237" }, { "image": "data/prediction/urban_day_rittenhouse/image/000238.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000238.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 441.5, "y1": 193.0, "x2": 555.0, "y2": 269.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000238" }, { "image": "data/prediction/urban_day_rittenhouse/image/000238.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000238.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.5, "y1": 193.0, "x2": 555.0, "y2": 269.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000238" }, { "image": "data/prediction/urban_day_rittenhouse/image/000239.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000239.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 366.0, "y1": 202.0, "x2": 428.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000239" }, { "image": "data/prediction/urban_day_rittenhouse/image/000239.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000239.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 366.0, "y1": 202.0, "x2": 428.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000239" }, { "image": "data/prediction/urban_day_rittenhouse/image/000240.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000240.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 446.5, "y1": 204.5, "x2": 541.5, "y2": 272.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000240" }, { "image": "data/prediction/urban_day_rittenhouse/image/000240.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000240.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 446.5, "y1": 204.5, "x2": 541.5, "y2": 272.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000240" }, { "image": "data/prediction/urban_day_rittenhouse/image/000241.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000241.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 396.5, "y1": 174.0, "x2": 458.5, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000241" }, { "image": "data/prediction/urban_day_rittenhouse/image/000241.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000241.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 396.5, "y1": 174.0, "x2": 458.5, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000241" }, { "image": "data/prediction/urban_day_rittenhouse/image/000242.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000242.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.0, "y1": 163.0, "x2": 501.5, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000242" }, { "image": "data/prediction/urban_day_rittenhouse/image/000242.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000242.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.0, "y1": 163.0, "x2": 501.5, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000242" }, { "image": "data/prediction/urban_day_rittenhouse/image/000243.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000243.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 153.0, "x2": 581.5, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000243" }, { "image": "data/prediction/urban_day_rittenhouse/image/000243.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000243.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 153.0, "x2": 581.5, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000243" }, { "image": "data/prediction/urban_day_rittenhouse/image/000244.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000244.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 498.0, "y1": 118.5, "x2": 639.5, "y2": 349.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000244" }, { "image": "data/prediction/urban_day_rittenhouse/image/000244.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000244.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 498.0, "y1": 118.5, "x2": 639.5, "y2": 349.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000244" }, { "image": "data/prediction/urban_day_rittenhouse/image/000245.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000245.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 402.5, "y1": 178.5, "x2": 460.5, "y2": 220.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000245" }, { "image": "data/prediction/urban_day_rittenhouse/image/000245.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000245.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 402.5, "y1": 178.5, "x2": 460.5, "y2": 220.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000245" }, { "image": "data/prediction/urban_day_rittenhouse/image/000246.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000246.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 424.0, "y1": 176.5, "x2": 504.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000246" }, { "image": "data/prediction/urban_day_rittenhouse/image/000246.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000246.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.0, "y1": 176.5, "x2": 504.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000246" }, { "image": "data/prediction/urban_day_rittenhouse/image/000247.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000247.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.5, "y1": 168.0, "x2": 589.0, "y2": 261.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000247" }, { "image": "data/prediction/urban_day_rittenhouse/image/000247.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000247.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 168.0, "x2": 589.0, "y2": 261.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000247" }, { "image": "data/prediction/urban_day_rittenhouse/image/000248.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000248.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 507.5, "y1": 161.0, "x2": 639.5, "y2": 306.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000248" }, { "image": "data/prediction/urban_day_rittenhouse/image/000248.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000248.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 507.5, "y1": 161.0, "x2": 639.5, "y2": 306.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000248" }, { "image": "data/prediction/urban_day_rittenhouse/image/000249.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000249.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 331.0, "y1": 159.5, "x2": 386.5, "y2": 224.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000249" }, { "image": "data/prediction/urban_day_rittenhouse/image/000249.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000249.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.0, "y1": 159.5, "x2": 386.5, "y2": 224.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000249" }, { "image": "data/prediction/urban_day_rittenhouse/image/000250.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000250.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 339.0, "y1": 154.0, "x2": 401.5, "y2": 226.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000250" }, { "image": "data/prediction/urban_day_rittenhouse/image/000250.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000250.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 339.0, "y1": 154.0, "x2": 401.5, "y2": 226.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000250" }, { "image": "data/prediction/urban_day_rittenhouse/image/000251.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000251.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 338.0, "y1": 152.5, "x2": 409.0, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000251" }, { "image": "data/prediction/urban_day_rittenhouse/image/000251.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000251.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 338.0, "y1": 152.5, "x2": 409.0, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000251" }, { "image": "data/prediction/urban_day_rittenhouse/image/000252.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000252.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 430.0, "y1": 197.5, "x2": 492.0, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000252" }, { "image": "data/prediction/urban_day_rittenhouse/image/000252.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000252.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 430.0, "y1": 197.5, "x2": 492.0, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000252" }, { "image": "data/prediction/urban_day_rittenhouse/image/000253.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000253.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 199.5, "x2": 501.5, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000253" }, { "image": "data/prediction/urban_day_rittenhouse/image/000253.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000253.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 199.5, "x2": 501.5, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000253" }, { "image": "data/prediction/urban_day_rittenhouse/image/000254.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000254.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 201.0, "x2": 505.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000254" }, { "image": "data/prediction/urban_day_rittenhouse/image/000254.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000254.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 201.0, "x2": 505.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000254" }, { "image": "data/prediction/urban_day_rittenhouse/image/000255.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000255.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 201.0, "x2": 505.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000255" }, { "image": "data/prediction/urban_day_rittenhouse/image/000255.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000255.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 201.0, "x2": 505.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000255" }, { "image": "data/prediction/urban_day_rittenhouse/image/000256.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000256.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 201.0, "x2": 505.5, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000256" }, { "image": "data/prediction/urban_day_rittenhouse/image/000256.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000256.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 201.0, "x2": 505.5, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000256" }, { "image": "data/prediction/urban_day_rittenhouse/image/000257.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000257.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 201.5, "x2": 505.5, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000257" }, { "image": "data/prediction/urban_day_rittenhouse/image/000257.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000257.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 201.5, "x2": 505.5, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000257" }, { "image": "data/prediction/urban_day_rittenhouse/image/000258.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000258.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 201.5, "x2": 504.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000258" }, { "image": "data/prediction/urban_day_rittenhouse/image/000258.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000258.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 201.5, "x2": 504.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000258" }, { "image": "data/prediction/urban_day_rittenhouse/image/000259.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000259.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 201.5, "x2": 504.5, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000259" }, { "image": "data/prediction/urban_day_rittenhouse/image/000259.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000259.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 201.5, "x2": 504.5, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000259" }, { "image": "data/prediction/urban_day_rittenhouse/image/000260.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000260.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.5, "y1": 201.0, "x2": 505.5, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000260" }, { "image": "data/prediction/urban_day_rittenhouse/image/000260.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000260.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.5, "y1": 201.0, "x2": 505.5, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000260" }, { "image": "data/prediction/urban_day_rittenhouse/image/000261.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000261.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.0, "y1": 201.5, "x2": 504.0, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000261" }, { "image": "data/prediction/urban_day_rittenhouse/image/000261.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000261.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.0, "y1": 201.5, "x2": 504.0, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000261" }, { "image": "data/prediction/urban_day_rittenhouse/image/000262.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000262.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 338.0, "y1": 152.0, "x2": 416.0, "y2": 242.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000262" }, { "image": "data/prediction/urban_day_rittenhouse/image/000262.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000262.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 338.0, "y1": 152.0, "x2": 416.0, "y2": 242.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000262" }, { "image": "data/prediction/urban_day_rittenhouse/image/000263.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000263.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 332.5, "y1": 150.0, "x2": 418.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000263" }, { "image": "data/prediction/urban_day_rittenhouse/image/000263.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000263.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 332.5, "y1": 150.0, "x2": 418.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000263" }, { "image": "data/prediction/urban_day_rittenhouse/image/000264.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000264.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 337.5, "y1": 152.5, "x2": 423.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000264" }, { "image": "data/prediction/urban_day_rittenhouse/image/000264.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000264.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 337.5, "y1": 152.5, "x2": 423.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000264" }, { "image": "data/prediction/urban_day_rittenhouse/image/000265.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000265.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 333.5, "y1": 151.5, "x2": 419.5, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000265" }, { "image": "data/prediction/urban_day_rittenhouse/image/000265.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000265.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 333.5, "y1": 151.5, "x2": 419.5, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000265" }, { "image": "data/prediction/urban_day_rittenhouse/image/000266.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000266.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 332.5, "y1": 151.5, "x2": 419.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000266" }, { "image": "data/prediction/urban_day_rittenhouse/image/000266.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000266.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 332.5, "y1": 151.5, "x2": 419.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000266" }, { "image": "data/prediction/urban_day_rittenhouse/image/000267.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000267.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 332.0, "y1": 152.5, "x2": 417.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000267" }, { "image": "data/prediction/urban_day_rittenhouse/image/000267.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000267.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 332.0, "y1": 152.5, "x2": 417.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000267" }, { "image": "data/prediction/urban_day_rittenhouse/image/000268.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000268.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 330.5, "y1": 153.0, "x2": 417.0, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000268" }, { "image": "data/prediction/urban_day_rittenhouse/image/000268.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000268.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 330.5, "y1": 153.0, "x2": 417.0, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000268" }, { "image": "data/prediction/urban_day_rittenhouse/image/000269.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000269.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 333.0, "y1": 151.5, "x2": 421.0, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000269" }, { "image": "data/prediction/urban_day_rittenhouse/image/000269.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000269.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 333.0, "y1": 151.5, "x2": 421.0, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000269" }, { "image": "data/prediction/urban_day_rittenhouse/image/000270.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000270.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 333.5, "y1": 152.0, "x2": 419.0, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000270" }, { "image": "data/prediction/urban_day_rittenhouse/image/000270.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000270.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 333.5, "y1": 152.0, "x2": 419.0, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000270" }, { "image": "data/prediction/urban_day_rittenhouse/image/000271.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000271.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 333.0, "y1": 150.0, "x2": 419.5, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000271" }, { "image": "data/prediction/urban_day_rittenhouse/image/000271.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000271.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 333.0, "y1": 150.0, "x2": 419.5, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000271" }, { "image": "data/prediction/urban_day_rittenhouse/image/000272.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000272.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 331.5, "y1": 151.5, "x2": 414.0, "y2": 247.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000272" }, { "image": "data/prediction/urban_day_rittenhouse/image/000272.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000272.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.5, "y1": 151.5, "x2": 414.0, "y2": 247.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000272" }, { "image": "data/prediction/urban_day_rittenhouse/image/000273.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000273.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 334.5, "y1": 157.5, "x2": 412.0, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000273" }, { "image": "data/prediction/urban_day_rittenhouse/image/000273.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000273.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 334.5, "y1": 157.5, "x2": 412.0, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000273" }, { "image": "data/prediction/urban_day_rittenhouse/image/000274.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000274.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 327.0, "y1": 161.0, "x2": 398.5, "y2": 245.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000274" }, { "image": "data/prediction/urban_day_rittenhouse/image/000274.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000274.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 327.0, "y1": 161.0, "x2": 398.5, "y2": 245.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000274" }, { "image": "data/prediction/urban_day_rittenhouse/image/000275.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000275.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 331.5, "y1": 165.0, "x2": 397.5, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000275" }, { "image": "data/prediction/urban_day_rittenhouse/image/000275.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000275.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.5, "y1": 165.0, "x2": 397.5, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000275" }, { "image": "data/prediction/urban_day_rittenhouse/image/000276.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000276.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 318.5, "y1": 159.5, "x2": 383.5, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000276" }, { "image": "data/prediction/urban_day_rittenhouse/image/000276.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000276.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 318.5, "y1": 159.5, "x2": 383.5, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000276" }, { "image": "data/prediction/urban_day_rittenhouse/image/000277.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000277.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 314.0, "y1": 160.0, "x2": 373.5, "y2": 229.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000277" }, { "image": "data/prediction/urban_day_rittenhouse/image/000277.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000277.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 314.0, "y1": 160.0, "x2": 373.5, "y2": 229.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000277" }, { "image": "data/prediction/urban_day_rittenhouse/image/000278.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000278.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.0, "y1": 202.0, "x2": 504.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000278" }, { "image": "data/prediction/urban_day_rittenhouse/image/000278.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000278.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.0, "y1": 202.0, "x2": 504.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000278" }, { "image": "data/prediction/urban_day_rittenhouse/image/000279.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000279.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 201.5, "x2": 506.5, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000279" }, { "image": "data/prediction/urban_day_rittenhouse/image/000279.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000279.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 201.5, "x2": 506.5, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000279" }, { "image": "data/prediction/urban_day_rittenhouse/image/000280.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000280.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 438.0, "y1": 205.5, "x2": 512.5, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000280" }, { "image": "data/prediction/urban_day_rittenhouse/image/000280.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000280.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 438.0, "y1": 205.5, "x2": 512.5, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000280" }, { "image": "data/prediction/urban_day_rittenhouse/image/000281.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000281.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 445.5, "y1": 207.5, "x2": 527.0, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000281" }, { "image": "data/prediction/urban_day_rittenhouse/image/000281.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000281.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 445.5, "y1": 207.5, "x2": 527.0, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000281" }, { "image": "data/prediction/urban_day_rittenhouse/image/000282.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000282.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 448.0, "y1": 207.5, "x2": 551.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000282" }, { "image": "data/prediction/urban_day_rittenhouse/image/000282.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000282.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 448.0, "y1": 207.5, "x2": 551.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000282" }, { "image": "data/prediction/urban_day_rittenhouse/image/000283.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000283.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 477.5, "y1": 212.5, "x2": 617.5, "y2": 303.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000283" }, { "image": "data/prediction/urban_day_rittenhouse/image/000283.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000283.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 477.5, "y1": 212.5, "x2": 617.5, "y2": 303.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000283" }, { "image": "data/prediction/urban_day_rittenhouse/image/000284.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000284.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 425.5, "y1": 201.5, "x2": 477.5, "y2": 245.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000284" }, { "image": "data/prediction/urban_day_rittenhouse/image/000284.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000284.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 425.5, "y1": 201.5, "x2": 477.5, "y2": 245.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000284" }, { "image": "data/prediction/urban_day_rittenhouse/image/000285.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000285.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.5, "y1": 202.0, "x2": 484.5, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000285" }, { "image": "data/prediction/urban_day_rittenhouse/image/000285.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000285.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.5, "y1": 202.0, "x2": 484.5, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000285" }, { "image": "data/prediction/urban_day_rittenhouse/image/000286.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000286.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 431.0, "y1": 203.0, "x2": 517.5, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000286" }, { "image": "data/prediction/urban_day_rittenhouse/image/000286.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000286.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.0, "y1": 203.0, "x2": 517.5, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000286" }, { "image": "data/prediction/urban_day_rittenhouse/image/000287.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000287.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 474.5, "y1": 213.0, "x2": 600.5, "y2": 297.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000287" }, { "image": "data/prediction/urban_day_rittenhouse/image/000287.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000287.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 474.5, "y1": 213.0, "x2": 600.5, "y2": 297.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000287" }, { "image": "data/prediction/urban_day_rittenhouse/image/000288.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000288.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 402.5, "y1": 196.0, "x2": 458.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000288" }, { "image": "data/prediction/urban_day_rittenhouse/image/000288.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000288.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 402.5, "y1": 196.0, "x2": 458.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000288" }, { "image": "data/prediction/urban_day_rittenhouse/image/000289.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000289.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 414.0, "y1": 197.0, "x2": 493.0, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000289" }, { "image": "data/prediction/urban_day_rittenhouse/image/000289.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000289.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.0, "y1": 197.0, "x2": 493.0, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000289" }, { "image": "data/prediction/urban_day_rittenhouse/image/000290.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000290.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 446.5, "y1": 209.0, "x2": 551.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000290" }, { "image": "data/prediction/urban_day_rittenhouse/image/000290.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000290.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 446.5, "y1": 209.0, "x2": 551.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000290" }, { "image": "data/prediction/urban_day_rittenhouse/image/000291.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000291.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 484.0, "y1": 206.0, "x2": 639.5, "y2": 308.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000291" }, { "image": "data/prediction/urban_day_rittenhouse/image/000291.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000291.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 484.0, "y1": 206.0, "x2": 639.5, "y2": 308.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000291" }, { "image": "data/prediction/urban_day_rittenhouse/image/000292.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000292.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 323.5, "y1": 161.0, "x2": 382.0, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000292" }, { "image": "data/prediction/urban_day_rittenhouse/image/000292.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000292.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.5, "y1": 161.0, "x2": 382.0, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000292" }, { "image": "data/prediction/urban_day_rittenhouse/image/000293.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000293.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 323.0, "y1": 160.5, "x2": 382.0, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000293" }, { "image": "data/prediction/urban_day_rittenhouse/image/000293.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000293.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.0, "y1": 160.5, "x2": 382.0, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000293" }, { "image": "data/prediction/urban_day_rittenhouse/image/000294.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000294.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 323.5, "y1": 161.5, "x2": 381.5, "y2": 229.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000294" }, { "image": "data/prediction/urban_day_rittenhouse/image/000294.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000294.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.5, "y1": 161.5, "x2": 381.5, "y2": 229.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000294" }, { "image": "data/prediction/urban_day_rittenhouse/image/000295.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000295.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 491.5, "y1": 203.0, "x2": 581.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000295" }, { "image": "data/prediction/urban_day_rittenhouse/image/000295.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000295.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 491.5, "y1": 203.0, "x2": 581.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000295" }, { "image": "data/prediction/urban_day_rittenhouse/image/000296.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000296.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 532.5, "y1": 206.5, "x2": 639.5, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000296" }, { "image": "data/prediction/urban_day_rittenhouse/image/000296.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000296.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 532.5, "y1": 206.5, "x2": 639.5, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000296" }, { "image": "data/prediction/urban_day_rittenhouse/image/000297.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000297.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 416.5, "y1": 195.5, "x2": 475.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000297" }, { "image": "data/prediction/urban_day_rittenhouse/image/000297.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000297.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.5, "y1": 195.5, "x2": 475.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000297" }, { "image": "data/prediction/urban_day_rittenhouse/image/000298.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000298.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 437.5, "y1": 200.5, "x2": 522.0, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000298" }, { "image": "data/prediction/urban_day_rittenhouse/image/000298.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000298.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.5, "y1": 200.5, "x2": 522.0, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000298" }, { "image": "data/prediction/urban_day_rittenhouse/image/000299.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000299.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 468.0, "y1": 208.0, "x2": 600.0, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000299" }, { "image": "data/prediction/urban_day_rittenhouse/image/000299.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000299.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 468.0, "y1": 208.0, "x2": 600.0, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000299" }, { "image": "data/prediction/urban_day_rittenhouse/image/000300.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000300.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 323.0, "y1": 155.5, "x2": 388.0, "y2": 231.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000300" }, { "image": "data/prediction/urban_day_rittenhouse/image/000300.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000300.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.0, "y1": 155.5, "x2": 388.0, "y2": 231.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000300" }, { "image": "data/prediction/urban_day_rittenhouse/image/000301.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000301.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 318.5, "y1": 152.5, "x2": 386.5, "y2": 231.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000301" }, { "image": "data/prediction/urban_day_rittenhouse/image/000301.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000301.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 318.5, "y1": 152.5, "x2": 386.5, "y2": 231.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000301" }, { "image": "data/prediction/urban_day_rittenhouse/image/000302.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000302.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 326.0, "y1": 156.5, "x2": 396.5, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000302" }, { "image": "data/prediction/urban_day_rittenhouse/image/000302.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000302.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 326.0, "y1": 156.5, "x2": 396.5, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000302" }, { "image": "data/prediction/urban_day_rittenhouse/image/000303.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000303.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 316.0, "y1": 150.0, "x2": 400.5, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000303" }, { "image": "data/prediction/urban_day_rittenhouse/image/000303.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000303.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 316.0, "y1": 150.0, "x2": 400.5, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000303" }, { "image": "data/prediction/urban_day_rittenhouse/image/000304.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000304.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 294.5, "y1": 146.0, "x2": 418.5, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000304" }, { "image": "data/prediction/urban_day_rittenhouse/image/000304.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000304.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 294.5, "y1": 146.0, "x2": 418.5, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000304" }, { "image": "data/prediction/urban_day_rittenhouse/image/000305.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000305.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 236.5, "y1": 143.5, "x2": 397.5, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000305" }, { "image": "data/prediction/urban_day_rittenhouse/image/000305.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000305.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 236.5, "y1": 143.5, "x2": 397.5, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000305" }, { "image": "data/prediction/urban_day_rittenhouse/image/000306.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000306.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 175.5, "y1": 134.0, "x2": 371.0, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000306" }, { "image": "data/prediction/urban_day_rittenhouse/image/000306.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000306.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 175.5, "y1": 134.0, "x2": 371.0, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000306" }, { "image": "data/prediction/urban_day_rittenhouse/image/000307.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000307.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 69.0, "y1": 126.0, "x2": 301.0, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000307" }, { "image": "data/prediction/urban_day_rittenhouse/image/000307.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000307.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 69.0, "y1": 126.0, "x2": 301.0, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000307" }, { "image": "data/prediction/urban_day_rittenhouse/image/000308.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000308.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 121.0, "x2": 196.0, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000308" }, { "image": "data/prediction/urban_day_rittenhouse/image/000308.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000308.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 121.0, "x2": 196.0, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000308" }, { "image": "data/prediction/urban_day_rittenhouse/image/000309.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000309.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 423.5, "y1": 195.0, "x2": 475.0, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000309" }, { "image": "data/prediction/urban_day_rittenhouse/image/000309.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000309.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 423.5, "y1": 195.0, "x2": 475.0, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000309" }, { "image": "data/prediction/urban_day_rittenhouse/image/000310.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000310.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 438.5, "y1": 198.0, "x2": 514.5, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000310" }, { "image": "data/prediction/urban_day_rittenhouse/image/000310.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000310.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 438.5, "y1": 198.0, "x2": 514.5, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000310" }, { "image": "data/prediction/urban_day_rittenhouse/image/000311.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000311.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 467.0, "y1": 204.0, "x2": 582.0, "y2": 290.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000311" }, { "image": "data/prediction/urban_day_rittenhouse/image/000311.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000311.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 467.0, "y1": 204.0, "x2": 582.0, "y2": 290.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000311" }, { "image": "data/prediction/urban_day_rittenhouse/image/000312.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000312.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 517.0, "y1": 211.0, "x2": 639.5, "y2": 348.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000312" }, { "image": "data/prediction/urban_day_rittenhouse/image/000312.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000312.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 517.0, "y1": 211.0, "x2": 639.5, "y2": 348.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000312" }, { "image": "data/prediction/urban_day_rittenhouse/image/000313.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000313.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 34.5, "y1": 185.5, "x2": 94.5, "y2": 207.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000313" }, { "image": "data/prediction/urban_day_rittenhouse/image/000313.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000313.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 34.5, "y1": 185.5, "x2": 94.5, "y2": 207.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000313" }, { "image": "data/prediction/urban_day_rittenhouse/image/000314.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000314.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 402.0, "y1": 198.5, "x2": 462.0, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000314" }, { "image": "data/prediction/urban_day_rittenhouse/image/000314.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000314.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 402.0, "y1": 198.5, "x2": 462.0, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000314" }, { "image": "data/prediction/urban_day_rittenhouse/image/000315.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000315.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 418.0, "y1": 201.0, "x2": 490.5, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000315" }, { "image": "data/prediction/urban_day_rittenhouse/image/000315.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000315.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 418.0, "y1": 201.0, "x2": 490.5, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000315" }, { "image": "data/prediction/urban_day_rittenhouse/image/000316.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000316.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 433.0, "y1": 203.0, "x2": 533.5, "y2": 271.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000316" }, { "image": "data/prediction/urban_day_rittenhouse/image/000316.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000316.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 433.0, "y1": 203.0, "x2": 533.5, "y2": 271.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000316" }, { "image": "data/prediction/urban_day_rittenhouse/image/000317.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000317.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 456.5, "y1": 209.0, "x2": 597.0, "y2": 305.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000317" }, { "image": "data/prediction/urban_day_rittenhouse/image/000317.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000317.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 456.5, "y1": 209.0, "x2": 597.0, "y2": 305.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000317" }, { "image": "data/prediction/urban_day_rittenhouse/image/000318.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000318.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 500.5, "y1": 218.5, "x2": 639.5, "y2": 368.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000318" }, { "image": "data/prediction/urban_day_rittenhouse/image/000318.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000318.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 500.5, "y1": 218.5, "x2": 639.5, "y2": 368.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000318" }, { "image": "data/prediction/urban_day_rittenhouse/image/000319.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000319.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 181.5, "x2": 0.0, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000319" }, { "image": "data/prediction/urban_day_rittenhouse/image/000319.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000319.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 181.5, "x2": 0.0, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000319" }, { "image": "data/prediction/urban_day_rittenhouse/image/000320.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000320.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 34.5, "y1": 181.5, "x2": 154.0, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000320" }, { "image": "data/prediction/urban_day_rittenhouse/image/000320.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000320.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 34.5, "y1": 181.5, "x2": 154.0, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000320" }, { "image": "data/prediction/urban_day_rittenhouse/image/000321.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000321.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 171.0, "y1": 194.0, "x2": 320.0, "y2": 270.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000321" }, { "image": "data/prediction/urban_day_rittenhouse/image/000321.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000321.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 171.0, "y1": 194.0, "x2": 320.0, "y2": 270.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000321" }, { "image": "data/prediction/urban_day_rittenhouse/image/000322.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000322.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 299.0, "y1": 200.0, "x2": 480.0, "y2": 296.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000322" }, { "image": "data/prediction/urban_day_rittenhouse/image/000322.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000322.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.0, "y1": 200.0, "x2": 480.0, "y2": 296.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000322" }, { "image": "data/prediction/urban_day_rittenhouse/image/000323.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000323.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.0, "y1": 210.0, "x2": 639.5, "y2": 343.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000323" }, { "image": "data/prediction/urban_day_rittenhouse/image/000323.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000323.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.0, "y1": 210.0, "x2": 639.5, "y2": 343.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000323" }, { "image": "data/prediction/urban_day_rittenhouse/image/000324.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000324.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 107.5, "x2": 99.5, "y2": 264.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000324" }, { "image": "data/prediction/urban_day_rittenhouse/image/000324.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000324.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 107.5, "x2": 99.5, "y2": 264.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000324" }, { "image": "data/prediction/urban_day_rittenhouse/image/000325.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000325.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 4.0, "y1": 112.5, "x2": 168.5, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000325" }, { "image": "data/prediction/urban_day_rittenhouse/image/000325.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000325.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 4.0, "y1": 112.5, "x2": 168.5, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000325" }, { "image": "data/prediction/urban_day_rittenhouse/image/000326.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000326.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 87.0, "y1": 110.0, "x2": 232.5, "y2": 269.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000326" }, { "image": "data/prediction/urban_day_rittenhouse/image/000326.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000326.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 87.0, "y1": 110.0, "x2": 232.5, "y2": 269.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000326" }, { "image": "data/prediction/urban_day_rittenhouse/image/000327.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000327.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 169.5, "y1": 123.5, "x2": 302.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000327" }, { "image": "data/prediction/urban_day_rittenhouse/image/000327.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000327.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 169.5, "y1": 123.5, "x2": 302.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000327" }, { "image": "data/prediction/urban_day_rittenhouse/image/000328.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000328.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 196.0, "y1": 123.0, "x2": 345.5, "y2": 277.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000328" }, { "image": "data/prediction/urban_day_rittenhouse/image/000328.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000328.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 196.0, "y1": 123.0, "x2": 345.5, "y2": 277.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000328" }, { "image": "data/prediction/urban_day_rittenhouse/image/000329.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000329.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 252.0, "y1": 123.5, "x2": 456.0, "y2": 280.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000329" }, { "image": "data/prediction/urban_day_rittenhouse/image/000329.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000329.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 252.0, "y1": 123.5, "x2": 456.0, "y2": 280.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000329" }, { "image": "data/prediction/urban_day_rittenhouse/image/000330.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000330.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 326.5, "y1": 130.5, "x2": 585.5, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000330" }, { "image": "data/prediction/urban_day_rittenhouse/image/000330.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000330.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 326.5, "y1": 130.5, "x2": 585.5, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000330" }, { "image": "data/prediction/urban_day_rittenhouse/image/000331.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000331.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 64.0, "y1": 193.0, "x2": 150.0, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000331" }, { "image": "data/prediction/urban_day_rittenhouse/image/000331.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000331.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 64.0, "y1": 193.0, "x2": 150.0, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000331" }, { "image": "data/prediction/urban_day_rittenhouse/image/000332.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000332.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 179.0, "y1": 193.5, "x2": 278.5, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000332" }, { "image": "data/prediction/urban_day_rittenhouse/image/000332.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000332.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 179.0, "y1": 193.5, "x2": 278.5, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000332" }, { "image": "data/prediction/urban_day_rittenhouse/image/000333.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000333.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 290.0, "y1": 198.5, "x2": 403.5, "y2": 264.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000333" }, { "image": "data/prediction/urban_day_rittenhouse/image/000333.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000333.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 290.0, "y1": 198.5, "x2": 403.5, "y2": 264.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000333" }, { "image": "data/prediction/urban_day_rittenhouse/image/000334.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000334.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 389.5, "y1": 208.0, "x2": 525.0, "y2": 292.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000334" }, { "image": "data/prediction/urban_day_rittenhouse/image/000334.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000334.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 389.5, "y1": 208.0, "x2": 525.0, "y2": 292.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000334" }, { "image": "data/prediction/urban_day_rittenhouse/image/000335.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000335.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 449.5, "y1": 215.0, "x2": 614.5, "y2": 321.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000335" }, { "image": "data/prediction/urban_day_rittenhouse/image/000335.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000335.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 449.5, "y1": 215.0, "x2": 614.5, "y2": 321.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000335" }, { "image": "data/prediction/urban_day_rittenhouse/image/000336.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000336.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 522.5, "y1": 224.0, "x2": 639.5, "y2": 375.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000336" }, { "image": "data/prediction/urban_day_rittenhouse/image/000336.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000336.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 522.5, "y1": 224.0, "x2": 639.5, "y2": 375.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000336" }, { "image": "data/prediction/urban_day_rittenhouse/image/000337.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000337.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 212.0, "y1": 147.5, "x2": 330.5, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000337" }, { "image": "data/prediction/urban_day_rittenhouse/image/000337.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000337.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 212.0, "y1": 147.5, "x2": 330.5, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000337" }, { "image": "data/prediction/urban_day_rittenhouse/image/000338.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000338.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 409.0, "y1": 203.5, "x2": 466.5, "y2": 247.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000338" }, { "image": "data/prediction/urban_day_rittenhouse/image/000338.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000338.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 409.0, "y1": 203.5, "x2": 466.5, "y2": 247.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000338" }, { "image": "data/prediction/urban_day_rittenhouse/image/000339.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000339.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 424.5, "y1": 205.0, "x2": 507.5, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000339" }, { "image": "data/prediction/urban_day_rittenhouse/image/000339.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000339.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.5, "y1": 205.0, "x2": 507.5, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000339" }, { "image": "data/prediction/urban_day_rittenhouse/image/000340.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000340.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 460.5, "y1": 211.0, "x2": 595.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000340" }, { "image": "data/prediction/urban_day_rittenhouse/image/000340.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000340.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 460.5, "y1": 211.0, "x2": 595.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000340" }, { "image": "data/prediction/urban_day_rittenhouse/image/000341.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000341.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 394.0, "y1": 200.5, "x2": 446.0, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000341" }, { "image": "data/prediction/urban_day_rittenhouse/image/000341.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000341.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 394.0, "y1": 200.5, "x2": 446.0, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000341" }, { "image": "data/prediction/urban_day_rittenhouse/image/000342.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000342.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 421.0, "y1": 203.5, "x2": 499.0, "y2": 256.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000342" }, { "image": "data/prediction/urban_day_rittenhouse/image/000342.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000342.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.0, "y1": 203.5, "x2": 499.0, "y2": 256.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000342" }, { "image": "data/prediction/urban_day_rittenhouse/image/000343.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000343.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 399.0, "y1": 200.0, "x2": 452.5, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000343" }, { "image": "data/prediction/urban_day_rittenhouse/image/000343.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000343.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 399.0, "y1": 200.0, "x2": 452.5, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000343" }, { "image": "data/prediction/urban_day_rittenhouse/image/000344.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000344.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.5, "y1": 205.5, "x2": 512.5, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000344" }, { "image": "data/prediction/urban_day_rittenhouse/image/000344.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000344.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.5, "y1": 205.5, "x2": 512.5, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000344" }, { "image": "data/prediction/urban_day_rittenhouse/image/000345.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000345.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 411.5, "y1": 197.5, "x2": 461.5, "y2": 234.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000345" }, { "image": "data/prediction/urban_day_rittenhouse/image/000345.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000345.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.5, "y1": 197.5, "x2": 461.5, "y2": 234.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000345" }, { "image": "data/prediction/urban_day_rittenhouse/image/000346.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000346.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 439.5, "y1": 201.5, "x2": 524.5, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000346" }, { "image": "data/prediction/urban_day_rittenhouse/image/000346.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000346.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 439.5, "y1": 201.5, "x2": 524.5, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000346" }, { "image": "data/prediction/urban_day_rittenhouse/image/000347.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000347.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 403.0, "y1": 195.5, "x2": 458.5, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000347" }, { "image": "data/prediction/urban_day_rittenhouse/image/000347.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000347.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.0, "y1": 195.5, "x2": 458.5, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000347" }, { "image": "data/prediction/urban_day_rittenhouse/image/000348.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000348.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.0, "y1": 198.5, "x2": 515.0, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000348" }, { "image": "data/prediction/urban_day_rittenhouse/image/000348.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000348.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.0, "y1": 198.5, "x2": 515.0, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000348" }, { "image": "data/prediction/urban_day_rittenhouse/image/000349.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000349.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 480.0, "y1": 203.0, "x2": 639.5, "y2": 308.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000349" }, { "image": "data/prediction/urban_day_rittenhouse/image/000349.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000349.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 480.0, "y1": 203.0, "x2": 639.5, "y2": 308.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000349" }, { "image": "data/prediction/urban_day_rittenhouse/image/000350.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000350.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 280.5, "y1": 184.5, "x2": 334.0, "y2": 204.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000350" }, { "image": "data/prediction/urban_day_rittenhouse/image/000350.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000350.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.5, "y1": 184.5, "x2": 334.0, "y2": 204.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000350" }, { "image": "data/prediction/urban_day_rittenhouse/image/000351.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000351.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 266.0, "y1": 186.0, "x2": 334.0, "y2": 209.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000351" }, { "image": "data/prediction/urban_day_rittenhouse/image/000351.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000351.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 266.0, "y1": 186.0, "x2": 334.0, "y2": 209.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000351" }, { "image": "data/prediction/urban_day_rittenhouse/image/000352.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000352.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 243.5, "y1": 182.5, "x2": 329.5, "y2": 214.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000352" }, { "image": "data/prediction/urban_day_rittenhouse/image/000352.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000352.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.5, "y1": 182.5, "x2": 329.5, "y2": 214.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000352" }, { "image": "data/prediction/urban_day_rittenhouse/image/000353.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000353.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 226.5, "y1": 183.0, "x2": 332.5, "y2": 220.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000353" }, { "image": "data/prediction/urban_day_rittenhouse/image/000353.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000353.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 226.5, "y1": 183.0, "x2": 332.5, "y2": 220.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000353" }, { "image": "data/prediction/urban_day_rittenhouse/image/000354.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000354.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 205.5, "y1": 185.0, "x2": 336.5, "y2": 232.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000354" }, { "image": "data/prediction/urban_day_rittenhouse/image/000354.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000354.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 205.5, "y1": 185.0, "x2": 336.5, "y2": 232.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000354" }, { "image": "data/prediction/urban_day_rittenhouse/image/000355.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000355.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 184.0, "y1": 184.0, "x2": 348.5, "y2": 244.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000355" }, { "image": "data/prediction/urban_day_rittenhouse/image/000355.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000355.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 184.0, "y1": 184.0, "x2": 348.5, "y2": 244.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000355" }, { "image": "data/prediction/urban_day_rittenhouse/image/000356.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000356.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 179.0, "y1": 196.5, "x2": 384.5, "y2": 271.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000356" }, { "image": "data/prediction/urban_day_rittenhouse/image/000356.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000356.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 179.0, "y1": 196.5, "x2": 384.5, "y2": 271.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000356" }, { "image": "data/prediction/urban_day_rittenhouse/image/000357.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000357.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 64.0, "y1": 178.0, "x2": 142.5, "y2": 206.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000357" }, { "image": "data/prediction/urban_day_rittenhouse/image/000357.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000357.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 64.0, "y1": 178.0, "x2": 142.5, "y2": 206.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000357" }, { "image": "data/prediction/urban_day_rittenhouse/image/000358.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000358.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 10.5, "y1": 176.0, "x2": 107.5, "y2": 212.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000358" }, { "image": "data/prediction/urban_day_rittenhouse/image/000358.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000358.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 10.5, "y1": 176.0, "x2": 107.5, "y2": 212.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000358" }, { "image": "data/prediction/urban_day_rittenhouse/image/000359.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000359.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 177.5, "x2": 69.0, "y2": 220.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000359" }, { "image": "data/prediction/urban_day_rittenhouse/image/000359.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000359.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 177.5, "x2": 69.0, "y2": 220.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000359" }, { "image": "data/prediction/urban_day_rittenhouse/image/000360.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000360.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 182.0, "x2": 37.0, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000360" }, { "image": "data/prediction/urban_day_rittenhouse/image/000360.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000360.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 182.0, "x2": 37.0, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000360" }, { "image": "data/prediction/urban_day_rittenhouse/image/000361.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000361.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 196.5, "x2": 26.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000361" }, { "image": "data/prediction/urban_day_rittenhouse/image/000361.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000361.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 196.5, "x2": 26.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000361" }, { "image": "data/prediction/urban_day_rittenhouse/image/000362.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000362.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 200.5, "x2": 27.5, "y2": 272.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000362" }, { "image": "data/prediction/urban_day_rittenhouse/image/000362.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000362.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 200.5, "x2": 27.5, "y2": 272.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000362" }, { "image": "data/prediction/urban_day_rittenhouse/image/000363.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000363.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 205.0, "x2": 74.5, "y2": 287.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000363" }, { "image": "data/prediction/urban_day_rittenhouse/image/000363.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000363.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 205.0, "x2": 74.5, "y2": 287.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000363" }, { "image": "data/prediction/urban_day_rittenhouse/image/000364.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000364.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 49.0, "y1": 215.0, "x2": 256.5, "y2": 323.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000364" }, { "image": "data/prediction/urban_day_rittenhouse/image/000364.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000364.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 49.0, "y1": 215.0, "x2": 256.5, "y2": 323.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000364" }, { "image": "data/prediction/urban_day_rittenhouse/image/000365.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000365.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 305.0, "y1": 220.5, "x2": 610.5, "y2": 395.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000365" }, { "image": "data/prediction/urban_day_rittenhouse/image/000365.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000365.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.0, "y1": 220.5, "x2": 610.5, "y2": 395.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000365" }, { "image": "data/prediction/urban_day_rittenhouse/image/000366.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000366.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 199.0, "y1": 201.0, "x2": 474.0, "y2": 304.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000366" }, { "image": "data/prediction/urban_day_rittenhouse/image/000366.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000366.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 199.0, "y1": 201.0, "x2": 474.0, "y2": 304.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000366" }, { "image": "data/prediction/urban_day_rittenhouse/image/000367.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000367.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 279.5, "y1": 202.5, "x2": 626.0, "y2": 334.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000367" }, { "image": "data/prediction/urban_day_rittenhouse/image/000367.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000367.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 279.5, "y1": 202.5, "x2": 626.0, "y2": 334.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000367" }, { "image": "data/prediction/urban_day_rittenhouse/image/000368.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000368.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 205.5, "x2": 52.5, "y2": 277.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000368" }, { "image": "data/prediction/urban_day_rittenhouse/image/000368.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000368.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 205.5, "x2": 52.5, "y2": 277.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000368" }, { "image": "data/prediction/urban_day_rittenhouse/image/000369.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000369.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 149.0, "y1": 207.5, "x2": 301.5, "y2": 296.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000369" }, { "image": "data/prediction/urban_day_rittenhouse/image/000369.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000369.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 149.0, "y1": 207.5, "x2": 301.5, "y2": 296.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000369" }, { "image": "data/prediction/urban_day_rittenhouse/image/000370.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000370.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.0, "y1": 217.0, "x2": 639.5, "y2": 359.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000370" }, { "image": "data/prediction/urban_day_rittenhouse/image/000370.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000370.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.0, "y1": 217.0, "x2": 639.5, "y2": 359.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000370" }, { "image": "data/prediction/urban_day_rittenhouse/image/000371.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000371.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.5, "y1": 201.5, "x2": 475.0, "y2": 243.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000371" }, { "image": "data/prediction/urban_day_rittenhouse/image/000371.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000371.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.5, "y1": 201.5, "x2": 475.0, "y2": 243.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000371" }, { "image": "data/prediction/urban_day_rittenhouse/image/000372.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000372.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 462.5, "y1": 203.5, "x2": 534.0, "y2": 270.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000372" }, { "image": "data/prediction/urban_day_rittenhouse/image/000372.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000372.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 203.5, "x2": 534.0, "y2": 270.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000372" }, { "image": "data/prediction/urban_day_rittenhouse/image/000373.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000373.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 302.5, "y1": 180.0, "x2": 376.5, "y2": 209.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000373" }, { "image": "data/prediction/urban_day_rittenhouse/image/000373.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000373.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 302.5, "y1": 180.0, "x2": 376.5, "y2": 209.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000373" }, { "image": "data/prediction/urban_day_rittenhouse/image/000374.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000374.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 324.5, "y1": 195.0, "x2": 377.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000374" }, { "image": "data/prediction/urban_day_rittenhouse/image/000374.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000374.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.5, "y1": 195.0, "x2": 377.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000374" }, { "image": "data/prediction/urban_day_rittenhouse/image/000375.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000375.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 332.5, "y1": 193.0, "x2": 391.0, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000375" }, { "image": "data/prediction/urban_day_rittenhouse/image/000375.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000375.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 332.5, "y1": 193.0, "x2": 391.0, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000375" }, { "image": "data/prediction/urban_day_rittenhouse/image/000376.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000376.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 338.0, "y1": 197.0, "x2": 406.5, "y2": 257.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000376" }, { "image": "data/prediction/urban_day_rittenhouse/image/000376.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000376.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 338.0, "y1": 197.0, "x2": 406.5, "y2": 257.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000376" }, { "image": "data/prediction/urban_day_rittenhouse/image/000377.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000377.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 329.5, "y1": 202.5, "x2": 418.5, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000377" }, { "image": "data/prediction/urban_day_rittenhouse/image/000377.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000377.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 329.5, "y1": 202.5, "x2": 418.5, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000377" }, { "image": "data/prediction/urban_day_rittenhouse/image/000378.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000378.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 320.0, "y1": 206.5, "x2": 435.5, "y2": 310.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000378" }, { "image": "data/prediction/urban_day_rittenhouse/image/000378.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000378.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 320.0, "y1": 206.5, "x2": 435.5, "y2": 310.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000378" }, { "image": "data/prediction/urban_day_rittenhouse/image/000379.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000379.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 180.5, "y1": 183.5, "x2": 238.0, "y2": 204.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000379" }, { "image": "data/prediction/urban_day_rittenhouse/image/000379.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000379.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 180.5, "y1": 183.5, "x2": 238.0, "y2": 204.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000379" }, { "image": "data/prediction/urban_day_rittenhouse/image/000380.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000380.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 348.5, "y1": 158.0, "x2": 426.0, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000380" }, { "image": "data/prediction/urban_day_rittenhouse/image/000380.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000380.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 348.5, "y1": 158.0, "x2": 426.0, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000380" }, { "image": "data/prediction/urban_day_rittenhouse/image/000381.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000381.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 346.0, "y1": 152.0, "x2": 433.5, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000381" }, { "image": "data/prediction/urban_day_rittenhouse/image/000381.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000381.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 346.0, "y1": 152.0, "x2": 433.5, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000381" }, { "image": "data/prediction/urban_day_rittenhouse/image/000382.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000382.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.5, "y1": 200.5, "x2": 532.5, "y2": 260.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000382" }, { "image": "data/prediction/urban_day_rittenhouse/image/000382.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000382.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.5, "y1": 200.5, "x2": 532.5, "y2": 260.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000382" }, { "image": "data/prediction/urban_day_rittenhouse/image/000383.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000383.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 293.5, "y1": 194.0, "x2": 344.0, "y2": 224.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000383" }, { "image": "data/prediction/urban_day_rittenhouse/image/000383.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000383.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 293.5, "y1": 194.0, "x2": 344.0, "y2": 224.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000383" }, { "image": "data/prediction/urban_day_rittenhouse/image/000384.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000384.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 272.5, "y1": 192.0, "x2": 327.5, "y2": 227.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000384" }, { "image": "data/prediction/urban_day_rittenhouse/image/000384.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000384.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 272.5, "y1": 192.0, "x2": 327.5, "y2": 227.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000384" }, { "image": "data/prediction/urban_day_rittenhouse/image/000385.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000385.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 247.5, "y1": 193.0, "x2": 322.0, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000385" }, { "image": "data/prediction/urban_day_rittenhouse/image/000385.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000385.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 247.5, "y1": 193.0, "x2": 322.0, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000385" }, { "image": "data/prediction/urban_day_rittenhouse/image/000386.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000386.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 215.0, "y1": 190.0, "x2": 314.5, "y2": 242.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000386" }, { "image": "data/prediction/urban_day_rittenhouse/image/000386.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000386.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 215.0, "y1": 190.0, "x2": 314.5, "y2": 242.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000386" }, { "image": "data/prediction/urban_day_rittenhouse/image/000387.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000387.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 171.0, "y1": 189.0, "x2": 297.0, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000387" }, { "image": "data/prediction/urban_day_rittenhouse/image/000387.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000387.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 171.0, "y1": 189.0, "x2": 297.0, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000387" }, { "image": "data/prediction/urban_day_rittenhouse/image/000388.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000388.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 344.5, "y1": 149.5, "x2": 440.0, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000388" }, { "image": "data/prediction/urban_day_rittenhouse/image/000388.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000388.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 344.5, "y1": 149.5, "x2": 440.0, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000388" }, { "image": "data/prediction/urban_day_rittenhouse/image/000389.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000389.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 345.0, "y1": 152.5, "x2": 428.5, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000389" }, { "image": "data/prediction/urban_day_rittenhouse/image/000389.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000389.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 345.0, "y1": 152.5, "x2": 428.5, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000389" }, { "image": "data/prediction/urban_day_rittenhouse/image/000390.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000390.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 343.0, "y1": 160.5, "x2": 414.5, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000390" }, { "image": "data/prediction/urban_day_rittenhouse/image/000390.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000390.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 343.0, "y1": 160.5, "x2": 414.5, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000390" }, { "image": "data/prediction/urban_day_rittenhouse/image/000391.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000391.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 352.0, "y1": 166.0, "x2": 408.5, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000391" }, { "image": "data/prediction/urban_day_rittenhouse/image/000391.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000391.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 352.0, "y1": 166.0, "x2": 408.5, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000391" }, { "image": "data/prediction/urban_day_rittenhouse/image/000392.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000392.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 313.0, "y1": 211.5, "x2": 451.5, "y2": 337.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000392" }, { "image": "data/prediction/urban_day_rittenhouse/image/000392.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000392.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 313.0, "y1": 211.5, "x2": 451.5, "y2": 337.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000392" }, { "image": "data/prediction/urban_day_rittenhouse/image/000393.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000393.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 311.5, "y1": 215.0, "x2": 456.5, "y2": 347.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000393" }, { "image": "data/prediction/urban_day_rittenhouse/image/000393.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000393.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 311.5, "y1": 215.0, "x2": 456.5, "y2": 347.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000393" }, { "image": "data/prediction/urban_day_rittenhouse/image/000394.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000394.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 313.5, "y1": 213.0, "x2": 454.0, "y2": 341.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000394" }, { "image": "data/prediction/urban_day_rittenhouse/image/000394.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000394.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 313.5, "y1": 213.0, "x2": 454.0, "y2": 341.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000394" }, { "image": "data/prediction/urban_day_rittenhouse/image/000395.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000395.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 325.0, "y1": 211.5, "x2": 447.0, "y2": 323.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000395" }, { "image": "data/prediction/urban_day_rittenhouse/image/000395.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000395.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.0, "y1": 211.5, "x2": 447.0, "y2": 323.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000395" }, { "image": "data/prediction/urban_day_rittenhouse/image/000396.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000396.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 489.5, "y1": 200.5, "x2": 567.5, "y2": 275.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000396" }, { "image": "data/prediction/urban_day_rittenhouse/image/000396.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000396.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 489.5, "y1": 200.5, "x2": 567.5, "y2": 275.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000396" }, { "image": "data/prediction/urban_day_rittenhouse/image/000397.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000397.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 490.5, "y1": 201.0, "x2": 570.0, "y2": 278.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000397" }, { "image": "data/prediction/urban_day_rittenhouse/image/000397.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000397.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 490.5, "y1": 201.0, "x2": 570.0, "y2": 278.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000397" }, { "image": "data/prediction/urban_day_rittenhouse/image/000398.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000398.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 490.5, "y1": 201.5, "x2": 570.5, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000398" }, { "image": "data/prediction/urban_day_rittenhouse/image/000398.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000398.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 490.5, "y1": 201.5, "x2": 570.5, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000398" }, { "image": "data/prediction/urban_day_rittenhouse/image/000399.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000399.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 492.5, "y1": 204.0, "x2": 576.0, "y2": 283.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000399" }, { "image": "data/prediction/urban_day_rittenhouse/image/000399.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000399.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 492.5, "y1": 204.0, "x2": 576.0, "y2": 283.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000399" }, { "image": "data/prediction/urban_day_rittenhouse/image/000400.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000400.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 502.5, "y1": 206.5, "x2": 588.0, "y2": 291.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000400" }, { "image": "data/prediction/urban_day_rittenhouse/image/000400.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000400.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 502.5, "y1": 206.5, "x2": 588.0, "y2": 291.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000400" }, { "image": "data/prediction/urban_day_rittenhouse/image/000401.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000401.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 522.5, "y1": 209.0, "x2": 623.5, "y2": 307.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000401" }, { "image": "data/prediction/urban_day_rittenhouse/image/000401.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000401.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 522.5, "y1": 209.0, "x2": 623.5, "y2": 307.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000401" }, { "image": "data/prediction/urban_day_rittenhouse/image/000402.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000402.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 468.5, "y1": 200.5, "x2": 531.0, "y2": 255.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000402" }, { "image": "data/prediction/urban_day_rittenhouse/image/000402.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000402.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 468.5, "y1": 200.5, "x2": 531.0, "y2": 255.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000402" }, { "image": "data/prediction/urban_day_rittenhouse/image/000403.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000403.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.0, "y1": 202.0, "x2": 534.5, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000403" }, { "image": "data/prediction/urban_day_rittenhouse/image/000403.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000403.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.0, "y1": 202.0, "x2": 534.5, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000403" }, { "image": "data/prediction/urban_day_rittenhouse/image/000404.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000404.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 461.5, "y1": 204.0, "x2": 534.5, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000404" }, { "image": "data/prediction/urban_day_rittenhouse/image/000404.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000404.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 461.5, "y1": 204.0, "x2": 534.5, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000404" }, { "image": "data/prediction/urban_day_rittenhouse/image/000405.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000405.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 460.5, "y1": 204.0, "x2": 539.0, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000405" }, { "image": "data/prediction/urban_day_rittenhouse/image/000405.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000405.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 460.5, "y1": 204.0, "x2": 539.0, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000405" }, { "image": "data/prediction/urban_day_rittenhouse/image/000406.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000406.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.0, "y1": 208.0, "x2": 544.0, "y2": 268.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000406" }, { "image": "data/prediction/urban_day_rittenhouse/image/000406.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000406.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.0, "y1": 208.0, "x2": 544.0, "y2": 268.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000406" }, { "image": "data/prediction/urban_day_rittenhouse/image/000407.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000407.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 472.0, "y1": 208.0, "x2": 561.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000407" }, { "image": "data/prediction/urban_day_rittenhouse/image/000407.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000407.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 472.0, "y1": 208.0, "x2": 561.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000407" }, { "image": "data/prediction/urban_day_rittenhouse/image/000408.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000408.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 484.5, "y1": 211.5, "x2": 606.0, "y2": 297.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000408" }, { "image": "data/prediction/urban_day_rittenhouse/image/000408.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000408.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 484.5, "y1": 211.5, "x2": 606.0, "y2": 297.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000408" }, { "image": "data/prediction/urban_day_rittenhouse/image/000409.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000409.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 526.0, "y1": 218.0, "x2": 639.5, "y2": 344.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000409" }, { "image": "data/prediction/urban_day_rittenhouse/image/000409.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000409.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 526.0, "y1": 218.0, "x2": 639.5, "y2": 344.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000409" }, { "image": "data/prediction/urban_day_rittenhouse/image/000410.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000410.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 132.5, "y1": 186.0, "x2": 201.5, "y2": 210.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000410" }, { "image": "data/prediction/urban_day_rittenhouse/image/000410.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000410.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 132.5, "y1": 186.0, "x2": 201.5, "y2": 210.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000410" }, { "image": "data/prediction/urban_day_rittenhouse/image/000411.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000411.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 130.0, "y1": 186.5, "x2": 199.0, "y2": 211.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000411" }, { "image": "data/prediction/urban_day_rittenhouse/image/000411.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000411.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 130.0, "y1": 186.5, "x2": 199.0, "y2": 211.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000411" }, { "image": "data/prediction/urban_day_rittenhouse/image/000412.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000412.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 126.0, "y1": 188.5, "x2": 195.0, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000412" }, { "image": "data/prediction/urban_day_rittenhouse/image/000412.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000412.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 126.0, "y1": 188.5, "x2": 195.0, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000412" }, { "image": "data/prediction/urban_day_rittenhouse/image/000413.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000413.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 122.5, "y1": 189.5, "x2": 192.5, "y2": 214.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000413" }, { "image": "data/prediction/urban_day_rittenhouse/image/000413.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000413.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 122.5, "y1": 189.5, "x2": 192.5, "y2": 214.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000413" }, { "image": "data/prediction/urban_day_rittenhouse/image/000414.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000414.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 108.0, "y1": 193.5, "x2": 182.0, "y2": 218.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000414" }, { "image": "data/prediction/urban_day_rittenhouse/image/000414.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000414.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 108.0, "y1": 193.5, "x2": 182.0, "y2": 218.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000414" }, { "image": "data/prediction/urban_day_rittenhouse/image/000415.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000415.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 84.0, "y1": 192.0, "x2": 163.0, "y2": 219.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000415" }, { "image": "data/prediction/urban_day_rittenhouse/image/000415.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000415.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 84.0, "y1": 192.0, "x2": 163.0, "y2": 219.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000415" }, { "image": "data/prediction/urban_day_rittenhouse/image/000416.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000416.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 51.0, "y1": 189.5, "x2": 137.0, "y2": 220.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000416" }, { "image": "data/prediction/urban_day_rittenhouse/image/000416.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000416.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 51.0, "y1": 189.5, "x2": 137.0, "y2": 220.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000416" }, { "image": "data/prediction/urban_day_rittenhouse/image/000417.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000417.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 5.5, "y1": 183.5, "x2": 99.5, "y2": 218.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000417" }, { "image": "data/prediction/urban_day_rittenhouse/image/000417.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000417.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 5.5, "y1": 183.5, "x2": 99.5, "y2": 218.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000417" }, { "image": "data/prediction/urban_day_rittenhouse/image/000418.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000418.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 424.0, "y1": 204.5, "x2": 497.5, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000418" }, { "image": "data/prediction/urban_day_rittenhouse/image/000418.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000418.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.0, "y1": 204.5, "x2": 497.5, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000418" }, { "image": "data/prediction/urban_day_rittenhouse/image/000419.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000419.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 444.5, "y1": 209.5, "x2": 541.5, "y2": 275.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000419" }, { "image": "data/prediction/urban_day_rittenhouse/image/000419.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000419.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 444.5, "y1": 209.5, "x2": 541.5, "y2": 275.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000419" }, { "image": "data/prediction/urban_day_rittenhouse/image/000420.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000420.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 478.0, "y1": 214.0, "x2": 631.0, "y2": 310.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000420" }, { "image": "data/prediction/urban_day_rittenhouse/image/000420.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000420.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 478.0, "y1": 214.0, "x2": 631.0, "y2": 310.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000420" }, { "image": "data/prediction/urban_day_rittenhouse/image/000421.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000421.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 346.0, "y1": 202.5, "x2": 408.0, "y2": 254.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000421" }, { "image": "data/prediction/urban_day_rittenhouse/image/000421.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000421.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 346.0, "y1": 202.5, "x2": 408.0, "y2": 254.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000421" }, { "image": "data/prediction/urban_day_rittenhouse/image/000422.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000422.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 343.0, "y1": 193.0, "x2": 405.5, "y2": 242.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000422" }, { "image": "data/prediction/urban_day_rittenhouse/image/000422.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000422.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 343.0, "y1": 193.0, "x2": 405.5, "y2": 242.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000422" }, { "image": "data/prediction/urban_day_rittenhouse/image/000423.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000423.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 344.0, "y1": 199.5, "x2": 404.5, "y2": 247.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000423" }, { "image": "data/prediction/urban_day_rittenhouse/image/000423.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000423.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 344.0, "y1": 199.5, "x2": 404.5, "y2": 247.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000423" }, { "image": "data/prediction/urban_day_rittenhouse/image/000424.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000424.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 455.5, "y1": 203.0, "x2": 573.5, "y2": 283.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000424" }, { "image": "data/prediction/urban_day_rittenhouse/image/000424.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000424.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 455.5, "y1": 203.0, "x2": 573.5, "y2": 283.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000424" }, { "image": "data/prediction/urban_day_rittenhouse/image/000425.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000425.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 496.5, "y1": 207.5, "x2": 639.5, "y2": 329.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000425" }, { "image": "data/prediction/urban_day_rittenhouse/image/000425.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000425.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 496.5, "y1": 207.5, "x2": 639.5, "y2": 329.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000425" }, { "image": "data/prediction/urban_day_rittenhouse/image/000426.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000426.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 491.0, "y1": 203.5, "x2": 544.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000426" }, { "image": "data/prediction/urban_day_rittenhouse/image/000426.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000426.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 491.0, "y1": 203.5, "x2": 544.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000426" }, { "image": "data/prediction/urban_day_rittenhouse/image/000427.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000427.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 109.0, "y1": 190.0, "x2": 269.0, "y2": 286.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000427" }, { "image": "data/prediction/urban_day_rittenhouse/image/000427.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000427.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 109.0, "y1": 190.0, "x2": 269.0, "y2": 286.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000427" }, { "image": "data/prediction/urban_day_rittenhouse/image/000428.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000428.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 20.5, "y1": 198.0, "x2": 213.0, "y2": 323.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000428" }, { "image": "data/prediction/urban_day_rittenhouse/image/000428.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000428.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 20.5, "y1": 198.0, "x2": 213.0, "y2": 323.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000428" }, { "image": "data/prediction/urban_day_rittenhouse/image/000429.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000429.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 199.5, "x2": 135.0, "y2": 342.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000429" }, { "image": "data/prediction/urban_day_rittenhouse/image/000429.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000429.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 199.5, "x2": 135.0, "y2": 342.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000429" }, { "image": "data/prediction/urban_day_rittenhouse/image/000430.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000430.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 128.5, "x2": 19.5, "y2": 300.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000430" }, { "image": "data/prediction/urban_day_rittenhouse/image/000430.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000430.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 128.5, "x2": 19.5, "y2": 300.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000430" }, { "image": "data/prediction/urban_day_rittenhouse/image/000431.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000431.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 80.5, "x2": 162.0, "y2": 348.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000431" }, { "image": "data/prediction/urban_day_rittenhouse/image/000431.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000431.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 80.5, "x2": 162.0, "y2": 348.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000431" }, { "image": "data/prediction/urban_day_rittenhouse/image/000432.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000432.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 32.0, "y1": 204.0, "x2": 105.5, "y2": 261.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000432" }, { "image": "data/prediction/urban_day_rittenhouse/image/000432.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000432.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 32.0, "y1": 204.0, "x2": 105.5, "y2": 261.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000432" }, { "image": "data/prediction/urban_day_rittenhouse/image/000433.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000433.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 112.0, "y1": 206.5, "x2": 222.5, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000433" }, { "image": "data/prediction/urban_day_rittenhouse/image/000433.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000433.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 112.0, "y1": 206.5, "x2": 222.5, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000433" }, { "image": "data/prediction/urban_day_rittenhouse/image/000434.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000434.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 58.0, "y1": 214.5, "x2": 243.0, "y2": 348.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000434" }, { "image": "data/prediction/urban_day_rittenhouse/image/000434.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000434.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 58.0, "y1": 214.5, "x2": 243.0, "y2": 348.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000434" }, { "image": "data/prediction/urban_day_rittenhouse/image/000435.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000435.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 172.5, "y1": 205.5, "x2": 257.0, "y2": 277.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000435" }, { "image": "data/prediction/urban_day_rittenhouse/image/000435.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000435.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 172.5, "y1": 205.5, "x2": 257.0, "y2": 277.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000435" }, { "image": "data/prediction/urban_day_rittenhouse/image/000436.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000436.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 68.5, "y1": 215.0, "x2": 238.0, "y2": 334.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000436" }, { "image": "data/prediction/urban_day_rittenhouse/image/000436.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000436.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 68.5, "y1": 215.0, "x2": 238.0, "y2": 334.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000436" }, { "image": "data/prediction/urban_day_rittenhouse/image/000437.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000437.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 195.5, "y1": 146.5, "x2": 293.5, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000437" }, { "image": "data/prediction/urban_day_rittenhouse/image/000437.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000437.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 195.5, "y1": 146.5, "x2": 293.5, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000437" }, { "image": "data/prediction/urban_day_rittenhouse/image/000438.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000438.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 105.0, "y1": 120.5, "x2": 276.5, "y2": 291.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000438" }, { "image": "data/prediction/urban_day_rittenhouse/image/000438.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000438.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 105.0, "y1": 120.5, "x2": 276.5, "y2": 291.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000438" }, { "image": "data/prediction/urban_day_rittenhouse/image/000439.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000439.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 14.0, "x2": 238.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000439" }, { "image": "data/prediction/urban_day_rittenhouse/image/000439.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000439.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 14.0, "x2": 238.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000439" }, { "image": "data/prediction/urban_day_rittenhouse/image/000440.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000440.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 230.5, "y1": 189.5, "x2": 289.5, "y2": 242.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000440" }, { "image": "data/prediction/urban_day_rittenhouse/image/000440.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000440.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 230.5, "y1": 189.5, "x2": 289.5, "y2": 242.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000440" }, { "image": "data/prediction/urban_day_rittenhouse/image/000441.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000441.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 177.0, "y1": 200.5, "x2": 270.5, "y2": 272.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000441" }, { "image": "data/prediction/urban_day_rittenhouse/image/000441.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000441.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 177.0, "y1": 200.5, "x2": 270.5, "y2": 272.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000441" }, { "image": "data/prediction/urban_day_rittenhouse/image/000442.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000442.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 44.0, "y1": 200.0, "x2": 239.5, "y2": 345.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000442" }, { "image": "data/prediction/urban_day_rittenhouse/image/000442.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000442.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 44.0, "y1": 200.0, "x2": 239.5, "y2": 345.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000442" }, { "image": "data/prediction/urban_day_rittenhouse/image/000443.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000443.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 223.5, "y1": 186.0, "x2": 287.5, "y2": 243.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000443" }, { "image": "data/prediction/urban_day_rittenhouse/image/000443.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000443.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 223.5, "y1": 186.0, "x2": 287.5, "y2": 243.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000443" }, { "image": "data/prediction/urban_day_rittenhouse/image/000444.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000444.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 167.0, "y1": 179.5, "x2": 277.0, "y2": 272.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000444" }, { "image": "data/prediction/urban_day_rittenhouse/image/000444.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000444.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 167.0, "y1": 179.5, "x2": 277.0, "y2": 272.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000444" }, { "image": "data/prediction/urban_day_rittenhouse/image/000445.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000445.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 174.5, "x2": 243.0, "y2": 384.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000445" }, { "image": "data/prediction/urban_day_rittenhouse/image/000445.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000445.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 174.5, "x2": 243.0, "y2": 384.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000445" }, { "image": "data/prediction/urban_day_rittenhouse/image/000446.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000446.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 243.5, "y1": 190.5, "x2": 295.5, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000446" }, { "image": "data/prediction/urban_day_rittenhouse/image/000446.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000446.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.5, "y1": 190.5, "x2": 295.5, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000446" }, { "image": "data/prediction/urban_day_rittenhouse/image/000447.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000447.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 210.0, "y1": 186.5, "x2": 292.0, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000447" }, { "image": "data/prediction/urban_day_rittenhouse/image/000447.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000447.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 210.0, "y1": 186.5, "x2": 292.0, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000447" }, { "image": "data/prediction/urban_day_rittenhouse/image/000448.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000448.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 132.5, "y1": 186.0, "x2": 273.5, "y2": 293.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000448" }, { "image": "data/prediction/urban_day_rittenhouse/image/000448.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000448.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 132.5, "y1": 186.0, "x2": 273.5, "y2": 293.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000448" }, { "image": "data/prediction/urban_day_rittenhouse/image/000449.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000449.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 244.5, "y1": 200.0, "x2": 297.5, "y2": 243.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000449" }, { "image": "data/prediction/urban_day_rittenhouse/image/000449.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000449.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 244.5, "y1": 200.0, "x2": 297.5, "y2": 243.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000449" }, { "image": "data/prediction/urban_day_rittenhouse/image/000450.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000450.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 197.5, "y1": 204.5, "x2": 277.5, "y2": 268.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000450" }, { "image": "data/prediction/urban_day_rittenhouse/image/000450.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000450.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 197.5, "y1": 204.5, "x2": 277.5, "y2": 268.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000450" }, { "image": "data/prediction/urban_day_rittenhouse/image/000451.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000451.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 67.0, "y1": 212.0, "x2": 241.0, "y2": 341.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000451" }, { "image": "data/prediction/urban_day_rittenhouse/image/000451.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000451.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 67.0, "y1": 212.0, "x2": 241.0, "y2": 341.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000451" }, { "image": "data/prediction/urban_day_rittenhouse/image/000452.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000452.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 206.0, "y1": 187.0, "x2": 284.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000452" }, { "image": "data/prediction/urban_day_rittenhouse/image/000452.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000452.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 206.0, "y1": 187.0, "x2": 284.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000452" }, { "image": "data/prediction/urban_day_rittenhouse/image/000453.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000453.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 113.0, "y1": 188.0, "x2": 251.0, "y2": 294.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000453" }, { "image": "data/prediction/urban_day_rittenhouse/image/000453.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000453.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 113.0, "y1": 188.0, "x2": 251.0, "y2": 294.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000453" }, { "image": "data/prediction/urban_day_rittenhouse/image/000454.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000454.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 227.5, "y1": 199.0, "x2": 299.5, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000454" }, { "image": "data/prediction/urban_day_rittenhouse/image/000454.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000454.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 227.5, "y1": 199.0, "x2": 299.5, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000454" }, { "image": "data/prediction/urban_day_rittenhouse/image/000455.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000455.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 159.0, "y1": 201.5, "x2": 278.0, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000455" }, { "image": "data/prediction/urban_day_rittenhouse/image/000455.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000455.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 159.0, "y1": 201.5, "x2": 278.0, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000455" }, { "image": "data/prediction/urban_day_rittenhouse/image/000456.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000456.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 218.5, "y1": 199.0, "x2": 300.5, "y2": 260.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000456" }, { "image": "data/prediction/urban_day_rittenhouse/image/000456.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000456.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 218.5, "y1": 199.0, "x2": 300.5, "y2": 260.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000456" }, { "image": "data/prediction/urban_day_rittenhouse/image/000457.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000457.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 202.0, "y1": 197.0, "x2": 280.5, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000457" }, { "image": "data/prediction/urban_day_rittenhouse/image/000457.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000457.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 202.0, "y1": 197.0, "x2": 280.5, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000457" }, { "image": "data/prediction/urban_day_rittenhouse/image/000458.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000458.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 103.0, "y1": 200.0, "x2": 252.0, "y2": 314.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000458" }, { "image": "data/prediction/urban_day_rittenhouse/image/000458.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000458.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 103.0, "y1": 200.0, "x2": 252.0, "y2": 314.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000458" }, { "image": "data/prediction/urban_day_rittenhouse/image/000459.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000459.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 214.0, "y1": 194.5, "x2": 284.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000459" }, { "image": "data/prediction/urban_day_rittenhouse/image/000459.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000459.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 214.0, "y1": 194.5, "x2": 284.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000459" }, { "image": "data/prediction/urban_day_rittenhouse/image/000460.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000460.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 146.0, "y1": 198.5, "x2": 265.0, "y2": 287.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000460" }, { "image": "data/prediction/urban_day_rittenhouse/image/000460.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000460.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 146.0, "y1": 198.5, "x2": 265.0, "y2": 287.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000460" }, { "image": "data/prediction/urban_day_rittenhouse/image/000461.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000461.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 250.5, "y1": 192.5, "x2": 309.5, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000461" }, { "image": "data/prediction/urban_day_rittenhouse/image/000461.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000461.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 250.5, "y1": 192.5, "x2": 309.5, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000461" }, { "image": "data/prediction/urban_day_rittenhouse/image/000462.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000462.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 208.5, "y1": 198.5, "x2": 285.5, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000462" }, { "image": "data/prediction/urban_day_rittenhouse/image/000462.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000462.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 208.5, "y1": 198.5, "x2": 285.5, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000462" }, { "image": "data/prediction/urban_day_rittenhouse/image/000463.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000463.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 122.5, "y1": 204.5, "x2": 262.5, "y2": 304.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000463" }, { "image": "data/prediction/urban_day_rittenhouse/image/000463.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000463.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 122.5, "y1": 204.5, "x2": 262.5, "y2": 304.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000463" }, { "image": "data/prediction/urban_day_rittenhouse/image/000464.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000464.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 220.0, "x2": 224.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000464" }, { "image": "data/prediction/urban_day_rittenhouse/image/000464.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000464.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 220.0, "x2": 224.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000464" }, { "image": "data/prediction/urban_day_rittenhouse/image/000465.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000465.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 402.0, "y2": 207.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000465" }, { "image": "data/prediction/urban_day_rittenhouse/image/000465.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000465.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 402.0, "y2": 207.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000465" }, { "image": "data/prediction/urban_day_rittenhouse/image/000466.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000466.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 355.0, "y1": 155.0, "x2": 407.5, "y2": 211.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000466" }, { "image": "data/prediction/urban_day_rittenhouse/image/000466.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000466.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 355.0, "y1": 155.0, "x2": 407.5, "y2": 211.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000466" }, { "image": "data/prediction/urban_day_rittenhouse/image/000467.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000467.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 154.5, "x2": 408.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000467" }, { "image": "data/prediction/urban_day_rittenhouse/image/000467.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000467.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 154.5, "x2": 408.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000467" }, { "image": "data/prediction/urban_day_rittenhouse/image/000468.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000468.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 409.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000468" }, { "image": "data/prediction/urban_day_rittenhouse/image/000468.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000468.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 409.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000468" }, { "image": "data/prediction/urban_day_rittenhouse/image/000469.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000469.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 409.5, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000469" }, { "image": "data/prediction/urban_day_rittenhouse/image/000469.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000469.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 409.5, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000469" }, { "image": "data/prediction/urban_day_rittenhouse/image/000470.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000470.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 553.0, "y1": 200.5, "x2": 639.5, "y2": 264.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000470" }, { "image": "data/prediction/urban_day_rittenhouse/image/000470.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000470.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 553.0, "y1": 200.5, "x2": 639.5, "y2": 264.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000470" }, { "image": "data/prediction/urban_day_rittenhouse/image/000471.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000471.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 409.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000471" }, { "image": "data/prediction/urban_day_rittenhouse/image/000471.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000471.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 409.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000471" }, { "image": "data/prediction/urban_day_rittenhouse/image/000472.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000472.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 409.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000472" }, { "image": "data/prediction/urban_day_rittenhouse/image/000472.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000472.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 409.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000472" }, { "image": "data/prediction/urban_day_rittenhouse/image/000473.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000473.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 357.0, "y1": 155.0, "x2": 409.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000473" }, { "image": "data/prediction/urban_day_rittenhouse/image/000473.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000473.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.0, "y1": 155.0, "x2": 409.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000473" }, { "image": "data/prediction/urban_day_rittenhouse/image/000474.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000474.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 357.5, "y1": 151.5, "x2": 409.0, "y2": 209.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000474" }, { "image": "data/prediction/urban_day_rittenhouse/image/000474.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000474.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.5, "y1": 151.5, "x2": 409.0, "y2": 209.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000474" }, { "image": "data/prediction/urban_day_rittenhouse/image/000475.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000475.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 406.5, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000475" }, { "image": "data/prediction/urban_day_rittenhouse/image/000475.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000475.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 406.5, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000475" }, { "image": "data/prediction/urban_day_rittenhouse/image/000476.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000476.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 357.5, "y1": 154.5, "x2": 409.0, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000476" }, { "image": "data/prediction/urban_day_rittenhouse/image/000476.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000476.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.5, "y1": 154.5, "x2": 409.0, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000476" }, { "image": "data/prediction/urban_day_rittenhouse/image/000477.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000477.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 357.0, "y1": 154.0, "x2": 407.5, "y2": 210.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000477" }, { "image": "data/prediction/urban_day_rittenhouse/image/000477.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000477.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.0, "y1": 154.0, "x2": 407.5, "y2": 210.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000477" }, { "image": "data/prediction/urban_day_rittenhouse/image/000478.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000478.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 355.5, "y1": 155.0, "x2": 407.0, "y2": 211.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000478" }, { "image": "data/prediction/urban_day_rittenhouse/image/000478.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000478.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 355.5, "y1": 155.0, "x2": 407.0, "y2": 211.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000478" }, { "image": "data/prediction/urban_day_rittenhouse/image/000479.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000479.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000479" }, { "image": "data/prediction/urban_day_rittenhouse/image/000479.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000479.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000479" }, { "image": "data/prediction/urban_day_rittenhouse/image/000480.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000480.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 356.0, "y1": 154.5, "x2": 409.5, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000480" }, { "image": "data/prediction/urban_day_rittenhouse/image/000480.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000480.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 154.5, "x2": 409.5, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000480" }, { "image": "data/prediction/urban_day_rittenhouse/image/000481.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000481.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 414.0, "y1": 200.5, "x2": 572.5, "y2": 259.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000481" }, { "image": "data/prediction/urban_day_rittenhouse/image/000481.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000481.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.0, "y1": 200.5, "x2": 572.5, "y2": 259.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000481" }, { "image": "data/prediction/urban_day_rittenhouse/image/000482.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000482.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 353.5, "y1": 200.0, "x2": 506.0, "y2": 257.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000482" }, { "image": "data/prediction/urban_day_rittenhouse/image/000482.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000482.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 353.5, "y1": 200.0, "x2": 506.0, "y2": 257.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000482" }, { "image": "data/prediction/urban_day_rittenhouse/image/000483.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000483.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 326.0, "y1": 198.5, "x2": 454.0, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000483" }, { "image": "data/prediction/urban_day_rittenhouse/image/000483.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000483.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 326.0, "y1": 198.5, "x2": 454.0, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000483" }, { "image": "data/prediction/urban_day_rittenhouse/image/000484.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000484.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 316.0, "y1": 196.5, "x2": 409.5, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000484" }, { "image": "data/prediction/urban_day_rittenhouse/image/000484.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000484.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 316.0, "y1": 196.5, "x2": 409.5, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000484" }, { "image": "data/prediction/urban_day_rittenhouse/image/000485.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000485.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 304.5, "y1": 191.5, "x2": 379.5, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000485" }, { "image": "data/prediction/urban_day_rittenhouse/image/000485.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000485.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.5, "y1": 191.5, "x2": 379.5, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000485" }, { "image": "data/prediction/urban_day_rittenhouse/image/000486.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000486.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 587.5, "y1": 194.0, "x2": 639.5, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000486" }, { "image": "data/prediction/urban_day_rittenhouse/image/000486.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000486.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 587.5, "y1": 194.0, "x2": 639.5, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000486" }, { "image": "data/prediction/urban_day_rittenhouse/image/000487.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000487.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 498.5, "y1": 192.0, "x2": 639.5, "y2": 255.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000487" }, { "image": "data/prediction/urban_day_rittenhouse/image/000487.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000487.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 498.5, "y1": 192.0, "x2": 639.5, "y2": 255.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000487" }, { "image": "data/prediction/urban_day_rittenhouse/image/000488.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000488.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 405.0, "y1": 192.5, "x2": 564.5, "y2": 254.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000488" }, { "image": "data/prediction/urban_day_rittenhouse/image/000488.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000488.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 405.0, "y1": 192.5, "x2": 564.5, "y2": 254.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000488" }, { "image": "data/prediction/urban_day_rittenhouse/image/000489.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000489.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 348.5, "y1": 191.5, "x2": 493.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000489" }, { "image": "data/prediction/urban_day_rittenhouse/image/000489.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000489.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 348.5, "y1": 191.5, "x2": 493.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000489" }, { "image": "data/prediction/urban_day_rittenhouse/image/000490.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000490.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 319.5, "y1": 191.0, "x2": 429.5, "y2": 242.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000490" }, { "image": "data/prediction/urban_day_rittenhouse/image/000490.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000490.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 319.5, "y1": 191.0, "x2": 429.5, "y2": 242.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000490" }, { "image": "data/prediction/urban_day_rittenhouse/image/000491.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000491.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 309.0, "y1": 189.5, "x2": 386.0, "y2": 234.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000491" }, { "image": "data/prediction/urban_day_rittenhouse/image/000491.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000491.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.0, "y1": 189.5, "x2": 386.0, "y2": 234.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000491" }, { "image": "data/prediction/urban_day_rittenhouse/image/000492.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000492.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 520.0, "y1": 202.5, "x2": 639.5, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000492" }, { "image": "data/prediction/urban_day_rittenhouse/image/000492.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000492.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 520.0, "y1": 202.5, "x2": 639.5, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000492" }, { "image": "data/prediction/urban_day_rittenhouse/image/000493.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000493.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 423.5, "y1": 202.5, "x2": 576.0, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000493" }, { "image": "data/prediction/urban_day_rittenhouse/image/000493.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000493.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 423.5, "y1": 202.5, "x2": 576.0, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000493" }, { "image": "data/prediction/urban_day_rittenhouse/image/000494.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000494.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 356.5, "y1": 198.0, "x2": 493.0, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000494" }, { "image": "data/prediction/urban_day_rittenhouse/image/000494.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000494.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 198.0, "x2": 493.0, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000494" }, { "image": "data/prediction/urban_day_rittenhouse/image/000495.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000495.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 328.5, "y1": 196.5, "x2": 429.5, "y2": 244.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000495" }, { "image": "data/prediction/urban_day_rittenhouse/image/000495.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000495.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 328.5, "y1": 196.5, "x2": 429.5, "y2": 244.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000495" }, { "image": "data/prediction/urban_day_rittenhouse/image/000496.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000496.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 357.5, "y1": 155.5, "x2": 406.5, "y2": 210.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000496" }, { "image": "data/prediction/urban_day_rittenhouse/image/000496.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000496.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.5, "y1": 155.5, "x2": 406.5, "y2": 210.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000496" }, { "image": "data/prediction/urban_day_rittenhouse/image/000497.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000497.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000497" }, { "image": "data/prediction/urban_day_rittenhouse/image/000497.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000497.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000497" }, { "image": "data/prediction/urban_day_rittenhouse/image/000498.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000498.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 357.0, "y1": 155.0, "x2": 409.5, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000498" }, { "image": "data/prediction/urban_day_rittenhouse/image/000498.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000498.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.0, "y1": 155.0, "x2": 409.5, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000498" }, { "image": "data/prediction/urban_day_rittenhouse/image/000499.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000499.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 356.5, "y1": 154.5, "x2": 408.5, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000499" }, { "image": "data/prediction/urban_day_rittenhouse/image/000499.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000499.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 154.5, "x2": 408.5, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000499" }, { "image": "data/prediction/urban_day_rittenhouse/image/000500.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000500.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 408.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000500" }, { "image": "data/prediction/urban_day_rittenhouse/image/000500.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000500.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 408.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000500" }, { "image": "data/prediction/urban_day_rittenhouse/image/000501.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000501.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.5, "y2": 214.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000501" }, { "image": "data/prediction/urban_day_rittenhouse/image/000501.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000501.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.5, "y2": 214.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000501" }, { "image": "data/prediction/urban_day_rittenhouse/image/000502.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000502.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 408.0, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000502" }, { "image": "data/prediction/urban_day_rittenhouse/image/000502.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000502.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.0, "x2": 408.0, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000502" }, { "image": "data/prediction/urban_day_rittenhouse/image/000503.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000503.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 358.0, "y1": 155.0, "x2": 407.5, "y2": 211.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000503" }, { "image": "data/prediction/urban_day_rittenhouse/image/000503.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000503.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 358.0, "y1": 155.0, "x2": 407.5, "y2": 211.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000503" }, { "image": "data/prediction/urban_day_rittenhouse/image/000504.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000504.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 356.5, "y1": 154.5, "x2": 408.0, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000504" }, { "image": "data/prediction/urban_day_rittenhouse/image/000504.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000504.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 154.5, "x2": 408.0, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000504" }, { "image": "data/prediction/urban_day_rittenhouse/image/000505.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000505.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 357.5, "y1": 155.0, "x2": 405.5, "y2": 206.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000505" }, { "image": "data/prediction/urban_day_rittenhouse/image/000505.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000505.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.5, "y1": 155.0, "x2": 405.5, "y2": 206.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000505" }, { "image": "data/prediction/urban_day_rittenhouse/image/000506.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000506.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 319.5, "y1": 195.5, "x2": 387.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000506" }, { "image": "data/prediction/urban_day_rittenhouse/image/000506.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000506.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 319.5, "y1": 195.5, "x2": 387.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000506" }, { "image": "data/prediction/urban_day_rittenhouse/image/000507.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000507.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 315.5, "y1": 192.0, "x2": 367.0, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000507" }, { "image": "data/prediction/urban_day_rittenhouse/image/000507.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000507.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 315.5, "y1": 192.0, "x2": 367.0, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000507" }, { "image": "data/prediction/urban_day_rittenhouse/image/000508.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000508.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 578.0, "y1": 199.5, "x2": 639.5, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000508" }, { "image": "data/prediction/urban_day_rittenhouse/image/000508.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000508.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 578.0, "y1": 199.5, "x2": 639.5, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000508" }, { "image": "data/prediction/urban_day_rittenhouse/image/000509.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000509.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 422.5, "y1": 203.0, "x2": 582.0, "y2": 264.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000509" }, { "image": "data/prediction/urban_day_rittenhouse/image/000509.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000509.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 422.5, "y1": 203.0, "x2": 582.0, "y2": 264.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000509" }, { "image": "data/prediction/urban_day_rittenhouse/image/000510.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000510.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 325.0, "y1": 197.5, "x2": 457.0, "y2": 254.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000510" }, { "image": "data/prediction/urban_day_rittenhouse/image/000510.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000510.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.0, "y1": 197.5, "x2": 457.0, "y2": 254.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000510" }, { "image": "data/prediction/urban_day_rittenhouse/image/000511.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000511.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 300.5, "y1": 195.0, "x2": 375.0, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000511" }, { "image": "data/prediction/urban_day_rittenhouse/image/000511.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000511.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 300.5, "y1": 195.0, "x2": 375.0, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000511" }, { "image": "data/prediction/urban_day_rittenhouse/image/000512.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000512.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 539.0, "y1": 199.5, "x2": 639.5, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000512" }, { "image": "data/prediction/urban_day_rittenhouse/image/000512.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000512.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 539.0, "y1": 199.5, "x2": 639.5, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000512" }, { "image": "data/prediction/urban_day_rittenhouse/image/000513.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000513.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 425.0, "y1": 197.5, "x2": 587.0, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000513" }, { "image": "data/prediction/urban_day_rittenhouse/image/000513.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000513.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 425.0, "y1": 197.5, "x2": 587.0, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000513" }, { "image": "data/prediction/urban_day_rittenhouse/image/000514.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000514.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 348.5, "y1": 196.0, "x2": 475.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000514" }, { "image": "data/prediction/urban_day_rittenhouse/image/000514.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000514.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 348.5, "y1": 196.0, "x2": 475.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000514" }, { "image": "data/prediction/urban_day_rittenhouse/image/000515.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000515.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 316.0, "y1": 194.0, "x2": 395.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000515" }, { "image": "data/prediction/urban_day_rittenhouse/image/000515.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000515.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 316.0, "y1": 194.0, "x2": 395.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000515" }, { "image": "data/prediction/urban_day_rittenhouse/image/000516.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000516.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 566.5, "y1": 191.5, "x2": 639.5, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000516" }, { "image": "data/prediction/urban_day_rittenhouse/image/000516.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000516.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 566.5, "y1": 191.5, "x2": 639.5, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000516" }, { "image": "data/prediction/urban_day_rittenhouse/image/000517.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000517.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 406.0, "y1": 187.5, "x2": 595.5, "y2": 257.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000517" }, { "image": "data/prediction/urban_day_rittenhouse/image/000517.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000517.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 406.0, "y1": 187.5, "x2": 595.5, "y2": 257.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000517" }, { "image": "data/prediction/urban_day_rittenhouse/image/000518.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000518.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 322.0, "y1": 186.5, "x2": 486.5, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000518" }, { "image": "data/prediction/urban_day_rittenhouse/image/000518.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000518.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 322.0, "y1": 186.5, "x2": 486.5, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000518" }, { "image": "data/prediction/urban_day_rittenhouse/image/000519.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000519.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 305.0, "y1": 187.5, "x2": 401.5, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000519" }, { "image": "data/prediction/urban_day_rittenhouse/image/000519.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000519.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.0, "y1": 187.5, "x2": 401.5, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000519" }, { "image": "data/prediction/urban_day_rittenhouse/image/000520.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000520.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 551.0, "y1": 202.5, "x2": 639.5, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000520" }, { "image": "data/prediction/urban_day_rittenhouse/image/000520.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000520.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 551.0, "y1": 202.5, "x2": 639.5, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000520" }, { "image": "data/prediction/urban_day_rittenhouse/image/000521.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000521.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 355.0, "y1": 155.0, "x2": 409.5, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000521" }, { "image": "data/prediction/urban_day_rittenhouse/image/000521.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000521.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 355.0, "y1": 155.0, "x2": 409.5, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000521" }, { "image": "data/prediction/urban_day_rittenhouse/image/000522.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000522.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 410.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000522" }, { "image": "data/prediction/urban_day_rittenhouse/image/000522.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000522.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 155.0, "x2": 410.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000522" }, { "image": "data/prediction/urban_day_rittenhouse/image/000523.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000523.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.5, "y1": 154.5, "x2": 407.5, "y2": 212.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000523" }, { "image": "data/prediction/urban_day_rittenhouse/image/000523.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000523.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 154.5, "x2": 407.5, "y2": 212.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000523" }, { "image": "data/prediction/urban_day_rittenhouse/image/000524.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000524.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.0, "y2": 213.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000524" }, { "image": "data/prediction/urban_day_rittenhouse/image/000524.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000524.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 155.5, "x2": 409.0, "y2": 213.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000524" }, { "image": "data/prediction/urban_day_rittenhouse/image/000525.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000525.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 154.5, "x2": 407.5, "y2": 213.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000525" }, { "image": "data/prediction/urban_day_rittenhouse/image/000525.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000525.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 154.5, "x2": 407.5, "y2": 213.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000525" }, { "image": "data/prediction/urban_day_rittenhouse/image/000526.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000526.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 357.0, "y1": 155.5, "x2": 410.0, "y2": 214.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000526" }, { "image": "data/prediction/urban_day_rittenhouse/image/000526.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000526.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 357.0, "y1": 155.5, "x2": 410.0, "y2": 214.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000526" }, { "image": "data/prediction/urban_day_rittenhouse/image/000527.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000527.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 356.5, "y1": 159.5, "x2": 410.5, "y2": 219.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000527" }, { "image": "data/prediction/urban_day_rittenhouse/image/000527.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000527.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 159.5, "x2": 410.5, "y2": 219.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000527" }, { "image": "data/prediction/urban_day_rittenhouse/image/000528.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000528.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 391.0, "y1": 163.0, "x2": 441.0, "y2": 223.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000528" }, { "image": "data/prediction/urban_day_rittenhouse/image/000528.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000528.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 391.0, "y1": 163.0, "x2": 441.0, "y2": 223.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000528" }, { "image": "data/prediction/urban_day_rittenhouse/image/000529.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000529.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 426.5, "y1": 160.0, "x2": 486.0, "y2": 225.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000529" }, { "image": "data/prediction/urban_day_rittenhouse/image/000529.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000529.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 426.5, "y1": 160.0, "x2": 486.0, "y2": 225.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000529" }, { "image": "data/prediction/urban_day_rittenhouse/image/000530.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000530.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 423.5, "y1": 159.5, "x2": 497.0, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000530" }, { "image": "data/prediction/urban_day_rittenhouse/image/000530.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000530.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 423.5, "y1": 159.5, "x2": 497.0, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000530" }, { "image": "data/prediction/urban_day_rittenhouse/image/000531.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000531.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 407.0, "y1": 203.0, "x2": 566.0, "y2": 261.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000531" }, { "image": "data/prediction/urban_day_rittenhouse/image/000531.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000531.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 407.0, "y1": 203.0, "x2": 566.0, "y2": 261.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000531" }, { "image": "data/prediction/urban_day_rittenhouse/image/000532.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000532.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 334.5, "y1": 200.0, "x2": 455.0, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000532" }, { "image": "data/prediction/urban_day_rittenhouse/image/000532.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000532.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 334.5, "y1": 200.0, "x2": 455.0, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000532" }, { "image": "data/prediction/urban_day_rittenhouse/image/000533.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000533.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 318.5, "y1": 195.0, "x2": 388.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000533" }, { "image": "data/prediction/urban_day_rittenhouse/image/000533.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000533.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 318.5, "y1": 195.0, "x2": 388.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000533" }, { "image": "data/prediction/urban_day_rittenhouse/image/000534.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000534.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 414.5, "y1": 152.5, "x2": 510.0, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000534" }, { "image": "data/prediction/urban_day_rittenhouse/image/000534.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000534.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.5, "y1": 152.5, "x2": 510.0, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000534" }, { "image": "data/prediction/urban_day_rittenhouse/image/000535.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000535.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 418.0, "y1": 128.0, "x2": 572.5, "y2": 269.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000535" }, { "image": "data/prediction/urban_day_rittenhouse/image/000535.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000535.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 418.0, "y1": 128.0, "x2": 572.5, "y2": 269.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000535" }, { "image": "data/prediction/urban_day_rittenhouse/image/000536.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000536.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 444.5, "y1": 56.0, "x2": 639.5, "y2": 354.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000536" }, { "image": "data/prediction/urban_day_rittenhouse/image/000536.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000536.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 444.5, "y1": 56.0, "x2": 639.5, "y2": 354.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000536" }, { "image": "data/prediction/urban_day_rittenhouse/image/000537.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000537.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 134.5, "y1": 193.0, "x2": 200.0, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000537" }, { "image": "data/prediction/urban_day_rittenhouse/image/000537.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000537.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 134.5, "y1": 193.0, "x2": 200.0, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000537" }, { "image": "data/prediction/urban_day_rittenhouse/image/000538.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000538.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 331.5, "y1": 204.0, "x2": 415.0, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000538" }, { "image": "data/prediction/urban_day_rittenhouse/image/000538.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000538.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.5, "y1": 204.0, "x2": 415.0, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000538" }, { "image": "data/prediction/urban_day_rittenhouse/image/000539.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000539.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 454.5, "y1": 214.0, "x2": 598.0, "y2": 319.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000539" }, { "image": "data/prediction/urban_day_rittenhouse/image/000539.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000539.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 454.5, "y1": 214.0, "x2": 598.0, "y2": 319.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000539" }, { "image": "data/prediction/urban_day_rittenhouse/image/000540.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000540.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 181.5, "x2": 69.0, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000540" }, { "image": "data/prediction/urban_day_rittenhouse/image/000540.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000540.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 181.5, "x2": 69.0, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000540" }, { "image": "data/prediction/urban_day_rittenhouse/image/000541.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000541.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 226.5, "y1": 178.5, "x2": 283.0, "y2": 217.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000541" }, { "image": "data/prediction/urban_day_rittenhouse/image/000541.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000541.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 226.5, "y1": 178.5, "x2": 283.0, "y2": 217.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000541" }, { "image": "data/prediction/urban_day_rittenhouse/image/000542.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000542.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 382.0, "y1": 188.0, "x2": 449.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000542" }, { "image": "data/prediction/urban_day_rittenhouse/image/000542.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000542.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 382.0, "y1": 188.0, "x2": 449.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000542" }, { "image": "data/prediction/urban_day_rittenhouse/image/000543.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000543.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 385.0, "y1": 204.0, "x2": 459.5, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000543" }, { "image": "data/prediction/urban_day_rittenhouse/image/000543.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000543.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 385.0, "y1": 204.0, "x2": 459.5, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000543" }, { "image": "data/prediction/urban_day_rittenhouse/image/000544.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000544.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 401.0, "y1": 172.5, "x2": 462.5, "y2": 232.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000544" }, { "image": "data/prediction/urban_day_rittenhouse/image/000544.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000544.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 401.0, "y1": 172.5, "x2": 462.5, "y2": 232.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000544" }, { "image": "data/prediction/urban_day_rittenhouse/image/000545.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000545.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.0, "y1": 160.0, "x2": 501.0, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000545" }, { "image": "data/prediction/urban_day_rittenhouse/image/000545.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000545.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.0, "y1": 160.0, "x2": 501.0, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000545" }, { "image": "data/prediction/urban_day_rittenhouse/image/000546.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000546.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 450.5, "y1": 159.0, "x2": 568.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000546" }, { "image": "data/prediction/urban_day_rittenhouse/image/000546.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000546.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 450.5, "y1": 159.0, "x2": 568.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000546" }, { "image": "data/prediction/urban_day_rittenhouse/image/000547.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000547.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.0, "y1": 152.0, "x2": 639.5, "y2": 300.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000547" }, { "image": "data/prediction/urban_day_rittenhouse/image/000547.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000547.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.0, "y1": 152.0, "x2": 639.5, "y2": 300.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000547" }, { "image": "data/prediction/urban_day_rittenhouse/image/000548.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000548.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 503.0, "y1": 116.5, "x2": 639.5, "y2": 367.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000548" }, { "image": "data/prediction/urban_day_rittenhouse/image/000548.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000548.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 503.0, "y1": 116.5, "x2": 639.5, "y2": 367.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000548" }, { "image": "data/prediction/urban_day_rittenhouse/image/000549.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000549.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.0, "y1": 193.5, "x2": 609.0, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000549" }, { "image": "data/prediction/urban_day_rittenhouse/image/000549.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000549.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.0, "y1": 193.5, "x2": 609.0, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000549" }, { "image": "data/prediction/urban_day_rittenhouse/image/000550.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000550.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 420.0, "y1": 197.0, "x2": 476.0, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000550" }, { "image": "data/prediction/urban_day_rittenhouse/image/000550.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000550.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 420.0, "y1": 197.0, "x2": 476.0, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000550" }, { "image": "data/prediction/urban_day_rittenhouse/image/000551.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000551.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 431.5, "y1": 201.5, "x2": 502.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000551" }, { "image": "data/prediction/urban_day_rittenhouse/image/000551.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000551.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.5, "y1": 201.5, "x2": 502.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000551" }, { "image": "data/prediction/urban_day_rittenhouse/image/000552.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000552.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 443.5, "y1": 202.0, "x2": 528.5, "y2": 260.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000552" }, { "image": "data/prediction/urban_day_rittenhouse/image/000552.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000552.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 443.5, "y1": 202.0, "x2": 528.5, "y2": 260.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000552" }, { "image": "data/prediction/urban_day_rittenhouse/image/000553.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000553.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 417.0, "y1": 180.0, "x2": 470.0, "y2": 216.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000553" }, { "image": "data/prediction/urban_day_rittenhouse/image/000553.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000553.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 417.0, "y1": 180.0, "x2": 470.0, "y2": 216.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000553" }, { "image": "data/prediction/urban_day_rittenhouse/image/000554.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000554.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 433.0, "y1": 185.0, "x2": 495.5, "y2": 227.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000554" }, { "image": "data/prediction/urban_day_rittenhouse/image/000554.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000554.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 433.0, "y1": 185.0, "x2": 495.5, "y2": 227.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000554" }, { "image": "data/prediction/urban_day_rittenhouse/image/000555.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000555.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 431.5, "y1": 180.5, "x2": 513.0, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000555" }, { "image": "data/prediction/urban_day_rittenhouse/image/000555.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000555.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.5, "y1": 180.5, "x2": 513.0, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000555" }, { "image": "data/prediction/urban_day_rittenhouse/image/000556.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000556.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.0, "y1": 168.5, "x2": 588.5, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000556" }, { "image": "data/prediction/urban_day_rittenhouse/image/000556.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000556.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.0, "y1": 168.5, "x2": 588.5, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000556" }, { "image": "data/prediction/urban_day_rittenhouse/image/000557.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000557.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 503.0, "y1": 161.5, "x2": 639.5, "y2": 317.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000557" }, { "image": "data/prediction/urban_day_rittenhouse/image/000557.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000557.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 503.0, "y1": 161.5, "x2": 639.5, "y2": 317.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000557" }, { "image": "data/prediction/urban_day_rittenhouse/image/000558.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000558.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 331.0, "y1": 197.0, "x2": 381.5, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000558" }, { "image": "data/prediction/urban_day_rittenhouse/image/000558.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000558.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.0, "y1": 197.0, "x2": 381.5, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000558" }, { "image": "data/prediction/urban_day_rittenhouse/image/000559.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000559.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 330.0, "y1": 198.0, "x2": 390.0, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000559" }, { "image": "data/prediction/urban_day_rittenhouse/image/000559.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000559.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 330.0, "y1": 198.0, "x2": 390.0, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000559" }, { "image": "data/prediction/urban_day_rittenhouse/image/000560.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000560.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 329.0, "y1": 187.5, "x2": 380.0, "y2": 233.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000560" }, { "image": "data/prediction/urban_day_rittenhouse/image/000560.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000560.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 329.0, "y1": 187.5, "x2": 380.0, "y2": 233.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000560" }, { "image": "data/prediction/urban_day_rittenhouse/image/000561.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000561.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.0, "y1": 192.0, "x2": 471.0, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000561" }, { "image": "data/prediction/urban_day_rittenhouse/image/000561.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000561.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.0, "y1": 192.0, "x2": 471.0, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000561" }, { "image": "data/prediction/urban_day_rittenhouse/image/000562.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000562.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 418.5, "y1": 190.5, "x2": 482.0, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000562" }, { "image": "data/prediction/urban_day_rittenhouse/image/000562.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000562.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 418.5, "y1": 190.5, "x2": 482.0, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000562" }, { "image": "data/prediction/urban_day_rittenhouse/image/000563.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000563.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 405.5, "y1": 192.0, "x2": 457.0, "y2": 229.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000563" }, { "image": "data/prediction/urban_day_rittenhouse/image/000563.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000563.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 405.5, "y1": 192.0, "x2": 457.0, "y2": 229.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000563" }, { "image": "data/prediction/urban_day_rittenhouse/image/000564.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000564.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 361.5, "y1": 170.5, "x2": 492.5, "y2": 215.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000564" }, { "image": "data/prediction/urban_day_rittenhouse/image/000564.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000564.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 361.5, "y1": 170.5, "x2": 492.5, "y2": 215.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000564" }, { "image": "data/prediction/urban_day_rittenhouse/image/000565.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000565.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 329.0, "y1": 170.5, "x2": 488.0, "y2": 215.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000565" }, { "image": "data/prediction/urban_day_rittenhouse/image/000565.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000565.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 329.0, "y1": 170.5, "x2": 488.0, "y2": 215.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000565" }, { "image": "data/prediction/urban_day_rittenhouse/image/000566.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000566.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 310.0, "y1": 169.5, "x2": 461.0, "y2": 215.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000566" }, { "image": "data/prediction/urban_day_rittenhouse/image/000566.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000566.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 310.0, "y1": 169.5, "x2": 461.0, "y2": 215.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000566" }, { "image": "data/prediction/urban_day_rittenhouse/image/000567.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000567.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 304.5, "y1": 160.0, "x2": 427.5, "y2": 204.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000567" }, { "image": "data/prediction/urban_day_rittenhouse/image/000567.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000567.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.5, "y1": 160.0, "x2": 427.5, "y2": 204.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000567" }, { "image": "data/prediction/urban_day_rittenhouse/image/000568.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000568.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 456.0, "y1": 204.0, "x2": 563.0, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000568" }, { "image": "data/prediction/urban_day_rittenhouse/image/000568.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000568.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 456.0, "y1": 204.0, "x2": 563.0, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000568" }, { "image": "data/prediction/urban_day_rittenhouse/image/000569.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000569.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.5, "y1": 205.5, "x2": 575.0, "y2": 278.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000569" }, { "image": "data/prediction/urban_day_rittenhouse/image/000569.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000569.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 205.5, "x2": 575.0, "y2": 278.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000569" }, { "image": "data/prediction/urban_day_rittenhouse/image/000570.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000570.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.0, "y1": 206.5, "x2": 578.5, "y2": 281.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000570" }, { "image": "data/prediction/urban_day_rittenhouse/image/000570.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000570.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.0, "y1": 206.5, "x2": 578.5, "y2": 281.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000570" }, { "image": "data/prediction/urban_day_rittenhouse/image/000571.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000571.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 466.0, "y1": 207.0, "x2": 581.0, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000571" }, { "image": "data/prediction/urban_day_rittenhouse/image/000571.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000571.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 466.0, "y1": 207.0, "x2": 581.0, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000571" }, { "image": "data/prediction/urban_day_rittenhouse/image/000572.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000572.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.5, "y1": 208.0, "x2": 581.5, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000572" }, { "image": "data/prediction/urban_day_rittenhouse/image/000572.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000572.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 208.0, "x2": 581.5, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000572" }, { "image": "data/prediction/urban_day_rittenhouse/image/000573.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000573.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.5, "y1": 207.5, "x2": 581.5, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000573" }, { "image": "data/prediction/urban_day_rittenhouse/image/000573.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000573.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.5, "y1": 207.5, "x2": 581.5, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000573" }, { "image": "data/prediction/urban_day_rittenhouse/image/000574.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000574.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.5, "y1": 207.5, "x2": 580.0, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000574" }, { "image": "data/prediction/urban_day_rittenhouse/image/000574.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000574.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 207.5, "x2": 580.0, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000574" }, { "image": "data/prediction/urban_day_rittenhouse/image/000575.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000575.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.5, "y1": 207.5, "x2": 582.5, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000575" }, { "image": "data/prediction/urban_day_rittenhouse/image/000575.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000575.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.5, "y1": 207.5, "x2": 582.5, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000575" }, { "image": "data/prediction/urban_day_rittenhouse/image/000576.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000576.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 466.5, "y1": 207.5, "x2": 581.0, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000576" }, { "image": "data/prediction/urban_day_rittenhouse/image/000576.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000576.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 466.5, "y1": 207.5, "x2": 581.0, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000576" }, { "image": "data/prediction/urban_day_rittenhouse/image/000577.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000577.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.5, "y1": 208.0, "x2": 582.0, "y2": 283.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000577" }, { "image": "data/prediction/urban_day_rittenhouse/image/000577.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000577.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 208.0, "x2": 582.0, "y2": 283.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000577" }, { "image": "data/prediction/urban_day_rittenhouse/image/000578.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000578.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 326.5, "y1": 185.5, "x2": 382.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000578" }, { "image": "data/prediction/urban_day_rittenhouse/image/000578.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000578.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 326.5, "y1": 185.5, "x2": 382.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000578" }, { "image": "data/prediction/urban_day_rittenhouse/image/000579.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000579.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 329.5, "y1": 186.0, "x2": 386.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000579" }, { "image": "data/prediction/urban_day_rittenhouse/image/000579.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000579.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 329.5, "y1": 186.0, "x2": 386.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000579" }, { "image": "data/prediction/urban_day_rittenhouse/image/000580.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000580.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 327.0, "y1": 183.5, "x2": 387.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000580" }, { "image": "data/prediction/urban_day_rittenhouse/image/000580.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000580.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 327.0, "y1": 183.5, "x2": 387.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000580" }, { "image": "data/prediction/urban_day_rittenhouse/image/000581.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000581.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 325.5, "y1": 183.5, "x2": 385.0, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000581" }, { "image": "data/prediction/urban_day_rittenhouse/image/000581.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000581.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.5, "y1": 183.5, "x2": 385.0, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000581" }, { "image": "data/prediction/urban_day_rittenhouse/image/000582.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000582.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 323.0, "y1": 182.5, "x2": 388.0, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000582" }, { "image": "data/prediction/urban_day_rittenhouse/image/000582.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000582.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.0, "y1": 182.5, "x2": 388.0, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000582" }, { "image": "data/prediction/urban_day_rittenhouse/image/000583.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000583.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 323.5, "y1": 183.0, "x2": 387.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000583" }, { "image": "data/prediction/urban_day_rittenhouse/image/000583.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000583.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.5, "y1": 183.0, "x2": 387.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000583" }, { "image": "data/prediction/urban_day_rittenhouse/image/000584.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000584.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 321.5, "y1": 183.5, "x2": 388.0, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000584" }, { "image": "data/prediction/urban_day_rittenhouse/image/000584.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000584.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 321.5, "y1": 183.5, "x2": 388.0, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000584" }, { "image": "data/prediction/urban_day_rittenhouse/image/000585.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000585.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 321.5, "y1": 182.5, "x2": 388.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000585" }, { "image": "data/prediction/urban_day_rittenhouse/image/000585.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000585.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 321.5, "y1": 182.5, "x2": 388.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000585" }, { "image": "data/prediction/urban_day_rittenhouse/image/000586.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000586.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 323.0, "y1": 182.5, "x2": 388.0, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000586" }, { "image": "data/prediction/urban_day_rittenhouse/image/000586.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000586.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.0, "y1": 182.5, "x2": 388.0, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000586" }, { "image": "data/prediction/urban_day_rittenhouse/image/000587.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000587.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 326.0, "y1": 183.5, "x2": 385.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000587" }, { "image": "data/prediction/urban_day_rittenhouse/image/000587.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000587.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 326.0, "y1": 183.5, "x2": 385.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000587" }, { "image": "data/prediction/urban_day_rittenhouse/image/000588.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000588.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 424.5, "y1": 190.5, "x2": 503.0, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000588" }, { "image": "data/prediction/urban_day_rittenhouse/image/000588.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000588.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.5, "y1": 190.5, "x2": 503.0, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000588" }, { "image": "data/prediction/urban_day_rittenhouse/image/000589.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000589.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.5, "y1": 191.5, "x2": 506.0, "y2": 247.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000589" }, { "image": "data/prediction/urban_day_rittenhouse/image/000589.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000589.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.5, "y1": 191.5, "x2": 506.0, "y2": 247.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000589" }, { "image": "data/prediction/urban_day_rittenhouse/image/000590.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000590.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 426.5, "y1": 193.0, "x2": 509.5, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000590" }, { "image": "data/prediction/urban_day_rittenhouse/image/000590.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000590.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 426.5, "y1": 193.0, "x2": 509.5, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000590" }, { "image": "data/prediction/urban_day_rittenhouse/image/000591.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000591.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 429.0, "y1": 193.0, "x2": 509.5, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000591" }, { "image": "data/prediction/urban_day_rittenhouse/image/000591.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000591.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 429.0, "y1": 193.0, "x2": 509.5, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000591" }, { "image": "data/prediction/urban_day_rittenhouse/image/000592.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000592.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.5, "y1": 193.0, "x2": 511.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000592" }, { "image": "data/prediction/urban_day_rittenhouse/image/000592.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000592.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.5, "y1": 193.0, "x2": 511.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000592" }, { "image": "data/prediction/urban_day_rittenhouse/image/000593.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000593.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.0, "y1": 193.0, "x2": 512.5, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000593" }, { "image": "data/prediction/urban_day_rittenhouse/image/000593.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000593.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.0, "y1": 193.0, "x2": 512.5, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000593" }, { "image": "data/prediction/urban_day_rittenhouse/image/000594.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000594.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.0, "y1": 193.5, "x2": 512.0, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000594" }, { "image": "data/prediction/urban_day_rittenhouse/image/000594.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000594.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.0, "y1": 193.5, "x2": 512.0, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000594" }, { "image": "data/prediction/urban_day_rittenhouse/image/000595.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000595.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 426.5, "y1": 193.5, "x2": 511.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000595" }, { "image": "data/prediction/urban_day_rittenhouse/image/000595.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000595.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 426.5, "y1": 193.5, "x2": 511.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000595" }, { "image": "data/prediction/urban_day_rittenhouse/image/000596.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000596.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 426.5, "y1": 193.5, "x2": 512.0, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000596" }, { "image": "data/prediction/urban_day_rittenhouse/image/000596.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000596.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 426.5, "y1": 193.5, "x2": 512.0, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000596" }, { "image": "data/prediction/urban_day_rittenhouse/image/000597.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000597.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.0, "y1": 193.5, "x2": 511.0, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000597" }, { "image": "data/prediction/urban_day_rittenhouse/image/000597.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000597.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.0, "y1": 193.5, "x2": 511.0, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000597" }, { "image": "data/prediction/urban_day_rittenhouse/image/000598.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000598.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.0, "y1": 191.0, "x2": 465.0, "y2": 230.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000598" }, { "image": "data/prediction/urban_day_rittenhouse/image/000598.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000598.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.0, "y1": 191.0, "x2": 465.0, "y2": 230.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000598" }, { "image": "data/prediction/urban_day_rittenhouse/image/000599.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000599.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 404.5, "y1": 191.5, "x2": 469.0, "y2": 232.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000599" }, { "image": "data/prediction/urban_day_rittenhouse/image/000599.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000599.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 404.5, "y1": 191.5, "x2": 469.0, "y2": 232.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000599" }, { "image": "data/prediction/urban_day_rittenhouse/image/000600.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000600.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.0, "y1": 193.0, "x2": 469.5, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000600" }, { "image": "data/prediction/urban_day_rittenhouse/image/000600.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000600.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.0, "y1": 193.0, "x2": 469.5, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000600" }, { "image": "data/prediction/urban_day_rittenhouse/image/000601.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000601.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.5, "y1": 191.5, "x2": 472.0, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000601" }, { "image": "data/prediction/urban_day_rittenhouse/image/000601.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000601.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.5, "y1": 191.5, "x2": 472.0, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000601" }, { "image": "data/prediction/urban_day_rittenhouse/image/000602.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000602.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.0, "y1": 193.0, "x2": 470.5, "y2": 233.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000602" }, { "image": "data/prediction/urban_day_rittenhouse/image/000602.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000602.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.0, "y1": 193.0, "x2": 470.5, "y2": 233.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000602" }, { "image": "data/prediction/urban_day_rittenhouse/image/000603.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000603.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 409.0, "y1": 195.0, "x2": 470.5, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000603" }, { "image": "data/prediction/urban_day_rittenhouse/image/000603.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000603.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 409.0, "y1": 195.0, "x2": 470.5, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000603" }, { "image": "data/prediction/urban_day_rittenhouse/image/000604.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000604.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.0, "y1": 195.5, "x2": 469.0, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000604" }, { "image": "data/prediction/urban_day_rittenhouse/image/000604.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000604.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.0, "y1": 195.5, "x2": 469.0, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000604" }, { "image": "data/prediction/urban_day_rittenhouse/image/000605.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000605.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 406.5, "y1": 196.5, "x2": 469.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000605" }, { "image": "data/prediction/urban_day_rittenhouse/image/000605.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000605.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 406.5, "y1": 196.5, "x2": 469.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000605" }, { "image": "data/prediction/urban_day_rittenhouse/image/000606.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000606.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 407.5, "y1": 195.5, "x2": 469.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000606" }, { "image": "data/prediction/urban_day_rittenhouse/image/000606.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000606.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.5, "y1": 195.5, "x2": 469.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000606" }, { "image": "data/prediction/urban_day_rittenhouse/image/000607.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000607.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 406.5, "y1": 195.5, "x2": 470.0, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000607" }, { "image": "data/prediction/urban_day_rittenhouse/image/000607.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000607.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 406.5, "y1": 195.5, "x2": 470.0, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000607" }, { "image": "data/prediction/urban_day_rittenhouse/image/000608.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000608.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 326.0, "y1": 197.5, "x2": 397.0, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000608" }, { "image": "data/prediction/urban_day_rittenhouse/image/000608.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000608.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 326.0, "y1": 197.5, "x2": 397.0, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000608" }, { "image": "data/prediction/urban_day_rittenhouse/image/000609.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000609.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 324.5, "y1": 198.0, "x2": 398.5, "y2": 263.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000609" }, { "image": "data/prediction/urban_day_rittenhouse/image/000609.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000609.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.5, "y1": 198.0, "x2": 398.5, "y2": 263.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000609" }, { "image": "data/prediction/urban_day_rittenhouse/image/000610.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000610.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 325.5, "y1": 200.0, "x2": 401.0, "y2": 266.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000610" }, { "image": "data/prediction/urban_day_rittenhouse/image/000610.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000610.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.5, "y1": 200.0, "x2": 401.0, "y2": 266.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000610" }, { "image": "data/prediction/urban_day_rittenhouse/image/000611.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000611.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 323.5, "y1": 200.5, "x2": 400.0, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000611" }, { "image": "data/prediction/urban_day_rittenhouse/image/000611.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000611.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.5, "y1": 200.5, "x2": 400.0, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000611" }, { "image": "data/prediction/urban_day_rittenhouse/image/000612.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000612.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 324.5, "y1": 199.5, "x2": 401.0, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000612" }, { "image": "data/prediction/urban_day_rittenhouse/image/000612.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000612.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.5, "y1": 199.5, "x2": 401.0, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000612" }, { "image": "data/prediction/urban_day_rittenhouse/image/000613.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000613.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 324.5, "y1": 199.5, "x2": 401.0, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000613" }, { "image": "data/prediction/urban_day_rittenhouse/image/000613.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000613.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.5, "y1": 199.5, "x2": 401.0, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000613" }, { "image": "data/prediction/urban_day_rittenhouse/image/000614.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000614.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 325.0, "y1": 199.5, "x2": 401.5, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000614" }, { "image": "data/prediction/urban_day_rittenhouse/image/000614.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000614.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.0, "y1": 199.5, "x2": 401.5, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000614" }, { "image": "data/prediction/urban_day_rittenhouse/image/000615.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000615.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 324.0, "y1": 199.5, "x2": 401.0, "y2": 268.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000615" }, { "image": "data/prediction/urban_day_rittenhouse/image/000615.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000615.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.0, "y1": 199.5, "x2": 401.0, "y2": 268.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000615" }, { "image": "data/prediction/urban_day_rittenhouse/image/000616.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000616.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 325.5, "y1": 200.5, "x2": 401.5, "y2": 268.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000616" }, { "image": "data/prediction/urban_day_rittenhouse/image/000616.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000616.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.5, "y1": 200.5, "x2": 401.5, "y2": 268.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000616" }, { "image": "data/prediction/urban_day_rittenhouse/image/000617.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000617.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 323.5, "y1": 199.5, "x2": 401.0, "y2": 268.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000617" }, { "image": "data/prediction/urban_day_rittenhouse/image/000617.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000617.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.5, "y1": 199.5, "x2": 401.0, "y2": 268.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000617" }, { "image": "data/prediction/urban_day_rittenhouse/image/000618.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000618.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.0, "y1": 207.0, "x2": 581.0, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000618" }, { "image": "data/prediction/urban_day_rittenhouse/image/000618.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000618.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.0, "y1": 207.0, "x2": 581.0, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000618" }, { "image": "data/prediction/urban_day_rittenhouse/image/000619.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000619.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 466.0, "y1": 207.5, "x2": 582.0, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000619" }, { "image": "data/prediction/urban_day_rittenhouse/image/000619.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000619.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 466.0, "y1": 207.5, "x2": 582.0, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000619" }, { "image": "data/prediction/urban_day_rittenhouse/image/000620.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000620.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.5, "y1": 208.0, "x2": 582.5, "y2": 283.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000620" }, { "image": "data/prediction/urban_day_rittenhouse/image/000620.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000620.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 208.0, "x2": 582.5, "y2": 283.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000620" }, { "image": "data/prediction/urban_day_rittenhouse/image/000621.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000621.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 468.5, "y1": 211.0, "x2": 592.0, "y2": 292.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000621" }, { "image": "data/prediction/urban_day_rittenhouse/image/000621.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000621.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 468.5, "y1": 211.0, "x2": 592.0, "y2": 292.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000621" }, { "image": "data/prediction/urban_day_rittenhouse/image/000622.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000622.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 487.5, "y1": 216.5, "x2": 635.0, "y2": 312.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000622" }, { "image": "data/prediction/urban_day_rittenhouse/image/000622.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000622.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 487.5, "y1": 216.5, "x2": 635.0, "y2": 312.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000622" }, { "image": "data/prediction/urban_day_rittenhouse/image/000623.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000623.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 427.5, "y1": 193.0, "x2": 512.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000623" }, { "image": "data/prediction/urban_day_rittenhouse/image/000623.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000623.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.5, "y1": 193.0, "x2": 512.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000623" }, { "image": "data/prediction/urban_day_rittenhouse/image/000624.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000624.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.5, "y1": 194.0, "x2": 513.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000624" }, { "image": "data/prediction/urban_day_rittenhouse/image/000624.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000624.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.5, "y1": 194.0, "x2": 513.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000624" }, { "image": "data/prediction/urban_day_rittenhouse/image/000625.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000625.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.0, "y1": 193.0, "x2": 513.5, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000625" }, { "image": "data/prediction/urban_day_rittenhouse/image/000625.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000625.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.0, "y1": 193.0, "x2": 513.5, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000625" }, { "image": "data/prediction/urban_day_rittenhouse/image/000626.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000626.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 429.0, "y1": 195.0, "x2": 519.5, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000626" }, { "image": "data/prediction/urban_day_rittenhouse/image/000626.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000626.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 429.0, "y1": 195.0, "x2": 519.5, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000626" }, { "image": "data/prediction/urban_day_rittenhouse/image/000627.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000627.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 441.0, "y1": 200.5, "x2": 540.5, "y2": 266.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000627" }, { "image": "data/prediction/urban_day_rittenhouse/image/000627.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000627.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.0, "y1": 200.5, "x2": 540.5, "y2": 266.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000627" }, { "image": "data/prediction/urban_day_rittenhouse/image/000628.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000628.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 465.0, "y1": 212.0, "x2": 594.0, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000628" }, { "image": "data/prediction/urban_day_rittenhouse/image/000628.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000628.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.0, "y1": 212.0, "x2": 594.0, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000628" }, { "image": "data/prediction/urban_day_rittenhouse/image/000629.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000629.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 511.0, "y1": 212.5, "x2": 639.5, "y2": 344.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000629" }, { "image": "data/prediction/urban_day_rittenhouse/image/000629.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000629.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 511.0, "y1": 212.5, "x2": 639.5, "y2": 344.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000629" }, { "image": "data/prediction/urban_day_rittenhouse/image/000630.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000630.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.5, "y1": 193.5, "x2": 469.5, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000630" }, { "image": "data/prediction/urban_day_rittenhouse/image/000630.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000630.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.5, "y1": 193.5, "x2": 469.5, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000630" }, { "image": "data/prediction/urban_day_rittenhouse/image/000631.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000631.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.5, "y1": 196.0, "x2": 468.5, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000631" }, { "image": "data/prediction/urban_day_rittenhouse/image/000631.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000631.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.5, "y1": 196.0, "x2": 468.5, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000631" }, { "image": "data/prediction/urban_day_rittenhouse/image/000632.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000632.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.5, "y1": 194.0, "x2": 470.0, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000632" }, { "image": "data/prediction/urban_day_rittenhouse/image/000632.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000632.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.5, "y1": 194.0, "x2": 470.0, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000632" }, { "image": "data/prediction/urban_day_rittenhouse/image/000633.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000633.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 409.0, "y1": 197.0, "x2": 475.5, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000633" }, { "image": "data/prediction/urban_day_rittenhouse/image/000633.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000633.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 409.0, "y1": 197.0, "x2": 475.5, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000633" }, { "image": "data/prediction/urban_day_rittenhouse/image/000634.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000634.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 414.0, "y1": 201.5, "x2": 488.5, "y2": 247.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000634" }, { "image": "data/prediction/urban_day_rittenhouse/image/000634.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000634.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.0, "y1": 201.5, "x2": 488.5, "y2": 247.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000634" }, { "image": "data/prediction/urban_day_rittenhouse/image/000635.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000635.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 431.0, "y1": 207.5, "x2": 522.0, "y2": 265.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000635" }, { "image": "data/prediction/urban_day_rittenhouse/image/000635.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000635.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.0, "y1": 207.5, "x2": 522.0, "y2": 265.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000635" }, { "image": "data/prediction/urban_day_rittenhouse/image/000636.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000636.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 455.0, "y1": 208.5, "x2": 580.5, "y2": 287.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000636" }, { "image": "data/prediction/urban_day_rittenhouse/image/000636.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000636.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 455.0, "y1": 208.5, "x2": 580.5, "y2": 287.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000636" }, { "image": "data/prediction/urban_day_rittenhouse/image/000637.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000637.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 500.0, "y1": 208.5, "x2": 639.5, "y2": 343.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000637" }, { "image": "data/prediction/urban_day_rittenhouse/image/000637.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000637.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 500.0, "y1": 208.5, "x2": 639.5, "y2": 343.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000637" }, { "image": "data/prediction/urban_day_rittenhouse/image/000638.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000638.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 324.5, "y1": 199.5, "x2": 401.5, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000638" }, { "image": "data/prediction/urban_day_rittenhouse/image/000638.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000638.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.5, "y1": 199.5, "x2": 401.5, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000638" }, { "image": "data/prediction/urban_day_rittenhouse/image/000639.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000639.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 324.0, "y1": 200.5, "x2": 398.5, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000639" }, { "image": "data/prediction/urban_day_rittenhouse/image/000639.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000639.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.0, "y1": 200.5, "x2": 398.5, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000639" }, { "image": "data/prediction/urban_day_rittenhouse/image/000640.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000640.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 325.5, "y1": 200.0, "x2": 393.5, "y2": 260.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000640" }, { "image": "data/prediction/urban_day_rittenhouse/image/000640.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000640.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.5, "y1": 200.0, "x2": 393.5, "y2": 260.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000640" }, { "image": "data/prediction/urban_day_rittenhouse/image/000641.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000641.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 325.5, "y1": 200.5, "x2": 385.5, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000641" }, { "image": "data/prediction/urban_day_rittenhouse/image/000641.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000641.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.5, "y1": 200.5, "x2": 385.5, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000641" }, { "image": "data/prediction/urban_day_rittenhouse/image/000642.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000642.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 321.5, "y1": 199.0, "x2": 378.5, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000642" }, { "image": "data/prediction/urban_day_rittenhouse/image/000642.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000642.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 321.5, "y1": 199.0, "x2": 378.5, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000642" }, { "image": "data/prediction/urban_day_rittenhouse/image/000643.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000643.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 325.5, "y1": 184.5, "x2": 384.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000643" }, { "image": "data/prediction/urban_day_rittenhouse/image/000643.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000643.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 325.5, "y1": 184.5, "x2": 384.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000643" }, { "image": "data/prediction/urban_day_rittenhouse/image/000644.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000644.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 330.0, "y1": 186.0, "x2": 380.5, "y2": 232.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000644" }, { "image": "data/prediction/urban_day_rittenhouse/image/000644.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000644.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 330.0, "y1": 186.0, "x2": 380.5, "y2": 232.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000644" }, { "image": "data/prediction/urban_day_rittenhouse/image/000645.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000645.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 317.5, "y1": 199.5, "x2": 371.5, "y2": 245.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000645" }, { "image": "data/prediction/urban_day_rittenhouse/image/000645.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000645.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 317.5, "y1": 199.5, "x2": 371.5, "y2": 245.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000645" }, { "image": "data/prediction/urban_day_rittenhouse/image/000646.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000646.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 421.5, "y1": 195.5, "x2": 479.0, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000646" }, { "image": "data/prediction/urban_day_rittenhouse/image/000646.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000646.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.5, "y1": 195.5, "x2": 479.0, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000646" }, { "image": "data/prediction/urban_day_rittenhouse/image/000647.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000647.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 440.5, "y1": 200.5, "x2": 511.5, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000647" }, { "image": "data/prediction/urban_day_rittenhouse/image/000647.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000647.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 440.5, "y1": 200.5, "x2": 511.5, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000647" }, { "image": "data/prediction/urban_day_rittenhouse/image/000648.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000648.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.0, "y1": 204.0, "x2": 570.5, "y2": 274.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000648" }, { "image": "data/prediction/urban_day_rittenhouse/image/000648.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000648.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.0, "y1": 204.0, "x2": 570.5, "y2": 274.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000648" }, { "image": "data/prediction/urban_day_rittenhouse/image/000649.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000649.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 506.5, "y1": 212.5, "x2": 639.5, "y2": 317.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000649" }, { "image": "data/prediction/urban_day_rittenhouse/image/000649.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000649.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 506.5, "y1": 212.5, "x2": 639.5, "y2": 317.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000649" }, { "image": "data/prediction/urban_day_rittenhouse/image/000650.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000650.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 311.5, "y1": 195.0, "x2": 363.0, "y2": 228.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000650" }, { "image": "data/prediction/urban_day_rittenhouse/image/000650.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000650.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 311.5, "y1": 195.0, "x2": 363.0, "y2": 228.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000650" }, { "image": "data/prediction/urban_day_rittenhouse/image/000651.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000651.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 315.5, "y1": 199.0, "x2": 378.0, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000651" }, { "image": "data/prediction/urban_day_rittenhouse/image/000651.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000651.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 315.5, "y1": 199.0, "x2": 378.0, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000651" }, { "image": "data/prediction/urban_day_rittenhouse/image/000652.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000652.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 344.0, "y1": 199.5, "x2": 409.5, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000652" }, { "image": "data/prediction/urban_day_rittenhouse/image/000652.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000652.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 344.0, "y1": 199.5, "x2": 409.5, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000652" }, { "image": "data/prediction/urban_day_rittenhouse/image/000653.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000653.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 355.5, "y1": 202.5, "x2": 429.5, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000653" }, { "image": "data/prediction/urban_day_rittenhouse/image/000653.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000653.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 355.5, "y1": 202.5, "x2": 429.5, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000653" }, { "image": "data/prediction/urban_day_rittenhouse/image/000654.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000654.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 359.0, "y1": 203.0, "x2": 439.5, "y2": 276.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000654" }, { "image": "data/prediction/urban_day_rittenhouse/image/000654.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000654.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 359.0, "y1": 203.0, "x2": 439.5, "y2": 276.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000654" }, { "image": "data/prediction/urban_day_rittenhouse/image/000655.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000655.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 346.5, "y1": 201.5, "x2": 448.0, "y2": 280.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000655" }, { "image": "data/prediction/urban_day_rittenhouse/image/000655.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000655.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 346.5, "y1": 201.5, "x2": 448.0, "y2": 280.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000655" }, { "image": "data/prediction/urban_day_rittenhouse/image/000656.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000656.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 331.0, "y1": 205.0, "x2": 452.5, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000656" }, { "image": "data/prediction/urban_day_rittenhouse/image/000656.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000656.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.0, "y1": 205.0, "x2": 452.5, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000656" }, { "image": "data/prediction/urban_day_rittenhouse/image/000657.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000657.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 348.5, "y1": 209.0, "x2": 470.0, "y2": 319.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000657" }, { "image": "data/prediction/urban_day_rittenhouse/image/000657.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000657.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 348.5, "y1": 209.0, "x2": 470.0, "y2": 319.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000657" }, { "image": "data/prediction/urban_day_rittenhouse/image/000658.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000658.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 409.5, "y1": 198.5, "x2": 470.5, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000658" }, { "image": "data/prediction/urban_day_rittenhouse/image/000658.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000658.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 409.5, "y1": 198.5, "x2": 470.5, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000658" }, { "image": "data/prediction/urban_day_rittenhouse/image/000659.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000659.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 458.5, "y1": 203.5, "x2": 555.5, "y2": 275.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000659" }, { "image": "data/prediction/urban_day_rittenhouse/image/000659.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000659.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 458.5, "y1": 203.5, "x2": 555.5, "y2": 275.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000659" }, { "image": "data/prediction/urban_day_rittenhouse/image/000660.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000660.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 226.5, "y1": 187.5, "x2": 284.0, "y2": 209.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000660" }, { "image": "data/prediction/urban_day_rittenhouse/image/000660.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000660.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 226.5, "y1": 187.5, "x2": 284.0, "y2": 209.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000660" }, { "image": "data/prediction/urban_day_rittenhouse/image/000661.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000661.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 432.0, "y1": 198.5, "x2": 495.5, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000661" }, { "image": "data/prediction/urban_day_rittenhouse/image/000661.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000661.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 432.0, "y1": 198.5, "x2": 495.5, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000661" }, { "image": "data/prediction/urban_day_rittenhouse/image/000662.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000662.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 478.0, "y1": 201.0, "x2": 573.0, "y2": 274.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000662" }, { "image": "data/prediction/urban_day_rittenhouse/image/000662.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000662.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 478.0, "y1": 201.0, "x2": 573.0, "y2": 274.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000662" }, { "image": "data/prediction/urban_day_rittenhouse/image/000663.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000663.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 71.0, "y1": 186.5, "x2": 144.0, "y2": 214.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000663" }, { "image": "data/prediction/urban_day_rittenhouse/image/000663.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000663.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 71.0, "y1": 186.5, "x2": 144.0, "y2": 214.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000663" }, { "image": "data/prediction/urban_day_rittenhouse/image/000664.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000664.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 90.5, "y1": 179.0, "x2": 187.5, "y2": 218.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000664" }, { "image": "data/prediction/urban_day_rittenhouse/image/000664.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000664.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 90.5, "y1": 179.0, "x2": 187.5, "y2": 218.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000664" }, { "image": "data/prediction/urban_day_rittenhouse/image/000665.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000665.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 180.5, "x2": 0.0, "y2": 230.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000665" }, { "image": "data/prediction/urban_day_rittenhouse/image/000665.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000665.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 180.5, "x2": 0.0, "y2": 230.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000665" }, { "image": "data/prediction/urban_day_rittenhouse/image/000666.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000666.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 4.5, "y1": 185.5, "x2": 131.0, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000666" }, { "image": "data/prediction/urban_day_rittenhouse/image/000666.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000666.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 4.5, "y1": 185.5, "x2": 131.0, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000666" }, { "image": "data/prediction/urban_day_rittenhouse/image/000667.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000667.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 217.0, "y1": 199.0, "x2": 396.5, "y2": 294.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000667" }, { "image": "data/prediction/urban_day_rittenhouse/image/000667.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000667.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 217.0, "y1": 199.0, "x2": 396.5, "y2": 294.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000667" }, { "image": "data/prediction/urban_day_rittenhouse/image/000668.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000668.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 42.5, "y1": 176.5, "x2": 99.5, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000668" }, { "image": "data/prediction/urban_day_rittenhouse/image/000668.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000668.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 42.5, "y1": 176.5, "x2": 99.5, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000668" }, { "image": "data/prediction/urban_day_rittenhouse/image/000669.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000669.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 235.5, "y1": 188.0, "x2": 322.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000669" }, { "image": "data/prediction/urban_day_rittenhouse/image/000669.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000669.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 235.5, "y1": 188.0, "x2": 322.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000669" }, { "image": "data/prediction/urban_day_rittenhouse/image/000670.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000670.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 425.5, "y1": 191.5, "x2": 550.5, "y2": 256.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000670" }, { "image": "data/prediction/urban_day_rittenhouse/image/000670.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000670.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 425.5, "y1": 191.5, "x2": 550.5, "y2": 256.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000670" }, { "image": "data/prediction/urban_day_rittenhouse/image/000671.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000671.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 142.5, "y1": 185.0, "x2": 225.5, "y2": 218.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000671" }, { "image": "data/prediction/urban_day_rittenhouse/image/000671.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000671.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 142.5, "y1": 185.0, "x2": 225.5, "y2": 218.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000671" }, { "image": "data/prediction/urban_day_rittenhouse/image/000672.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000672.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 135.5, "y1": 196.0, "x2": 236.0, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000672" }, { "image": "data/prediction/urban_day_rittenhouse/image/000672.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000672.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 135.5, "y1": 196.0, "x2": 236.0, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000672" }, { "image": "data/prediction/urban_day_rittenhouse/image/000673.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000673.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 342.5, "y1": 209.0, "x2": 479.0, "y2": 295.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000673" }, { "image": "data/prediction/urban_day_rittenhouse/image/000673.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000673.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 342.5, "y1": 209.0, "x2": 479.0, "y2": 295.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000673" }, { "image": "data/prediction/urban_day_rittenhouse/image/000674.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000674.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 416.0, "y1": 197.0, "x2": 479.0, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000674" }, { "image": "data/prediction/urban_day_rittenhouse/image/000674.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000674.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.0, "y1": 197.0, "x2": 479.0, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000674" }, { "image": "data/prediction/urban_day_rittenhouse/image/000675.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000675.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 201.0, "x2": 523.0, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000675" }, { "image": "data/prediction/urban_day_rittenhouse/image/000675.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000675.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 201.0, "x2": 523.0, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000675" }, { "image": "data/prediction/urban_day_rittenhouse/image/000676.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000676.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 470.0, "y1": 208.5, "x2": 597.5, "y2": 289.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000676" }, { "image": "data/prediction/urban_day_rittenhouse/image/000676.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000676.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 470.0, "y1": 208.5, "x2": 597.5, "y2": 289.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000676" }, { "image": "data/prediction/urban_day_rittenhouse/image/000677.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000677.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 523.0, "y1": 217.0, "x2": 639.5, "y2": 346.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000677" }, { "image": "data/prediction/urban_day_rittenhouse/image/000677.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000677.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 523.0, "y1": 217.0, "x2": 639.5, "y2": 346.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000677" }, { "image": "data/prediction/urban_day_rittenhouse/image/000678.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000678.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 417.0, "y1": 198.5, "x2": 481.0, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000678" }, { "image": "data/prediction/urban_day_rittenhouse/image/000678.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000678.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 417.0, "y1": 198.5, "x2": 481.0, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000678" }, { "image": "data/prediction/urban_day_rittenhouse/image/000679.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000679.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.5, "y1": 202.5, "x2": 519.5, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000679" }, { "image": "data/prediction/urban_day_rittenhouse/image/000679.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000679.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.5, "y1": 202.5, "x2": 519.5, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000679" }, { "image": "data/prediction/urban_day_rittenhouse/image/000680.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000680.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 472.0, "y1": 206.5, "x2": 590.0, "y2": 284.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000680" }, { "image": "data/prediction/urban_day_rittenhouse/image/000680.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000680.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 472.0, "y1": 206.5, "x2": 590.0, "y2": 284.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000680" }, { "image": "data/prediction/urban_day_rittenhouse/image/000681.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000681.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 528.5, "y1": 216.5, "x2": 639.5, "y2": 343.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000681" }, { "image": "data/prediction/urban_day_rittenhouse/image/000681.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000681.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 528.5, "y1": 216.5, "x2": 639.5, "y2": 343.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000681" }, { "image": "data/prediction/urban_day_rittenhouse/image/000682.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000682.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 429.0, "y1": 200.0, "x2": 492.5, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000682" }, { "image": "data/prediction/urban_day_rittenhouse/image/000682.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000682.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 429.0, "y1": 200.0, "x2": 492.5, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000682" }, { "image": "data/prediction/urban_day_rittenhouse/image/000683.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000683.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 444.0, "y1": 201.0, "x2": 526.0, "y2": 257.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000683" }, { "image": "data/prediction/urban_day_rittenhouse/image/000683.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000683.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 444.0, "y1": 201.0, "x2": 526.0, "y2": 257.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000683" }, { "image": "data/prediction/urban_day_rittenhouse/image/000684.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000684.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 470.5, "y1": 204.0, "x2": 577.0, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000684" }, { "image": "data/prediction/urban_day_rittenhouse/image/000684.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000684.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 470.5, "y1": 204.0, "x2": 577.0, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000684" }, { "image": "data/prediction/urban_day_rittenhouse/image/000685.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000685.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 502.0, "y1": 209.0, "x2": 639.5, "y2": 307.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000685" }, { "image": "data/prediction/urban_day_rittenhouse/image/000685.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000685.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 502.0, "y1": 209.0, "x2": 639.5, "y2": 307.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000685" }, { "image": "data/prediction/urban_day_rittenhouse/image/000686.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000686.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 305.5, "y1": 169.0, "x2": 360.0, "y2": 219.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000686" }, { "image": "data/prediction/urban_day_rittenhouse/image/000686.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000686.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.5, "y1": 169.0, "x2": 360.0, "y2": 219.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000686" }, { "image": "data/prediction/urban_day_rittenhouse/image/000687.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000687.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 291.5, "y1": 168.5, "x2": 354.0, "y2": 219.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000687" }, { "image": "data/prediction/urban_day_rittenhouse/image/000687.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000687.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 291.5, "y1": 168.5, "x2": 354.0, "y2": 219.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000687" }, { "image": "data/prediction/urban_day_rittenhouse/image/000688.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000688.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 258.5, "y1": 169.5, "x2": 344.5, "y2": 225.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000688" }, { "image": "data/prediction/urban_day_rittenhouse/image/000688.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000688.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 258.5, "y1": 169.5, "x2": 344.5, "y2": 225.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000688" }, { "image": "data/prediction/urban_day_rittenhouse/image/000689.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000689.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 214.5, "y1": 164.0, "x2": 324.5, "y2": 223.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000689" }, { "image": "data/prediction/urban_day_rittenhouse/image/000689.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000689.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 214.5, "y1": 164.0, "x2": 324.5, "y2": 223.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000689" }, { "image": "data/prediction/urban_day_rittenhouse/image/000690.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000690.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 88.0, "y1": 144.5, "x2": 271.0, "y2": 217.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000690" }, { "image": "data/prediction/urban_day_rittenhouse/image/000690.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000690.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 88.0, "y1": 144.5, "x2": 271.0, "y2": 217.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000690" }, { "image": "data/prediction/urban_day_rittenhouse/image/000691.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000691.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 13.5, "y1": 141.0, "x2": 189.5, "y2": 221.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000691" }, { "image": "data/prediction/urban_day_rittenhouse/image/000691.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000691.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 13.5, "y1": 141.0, "x2": 189.5, "y2": 221.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000691" }, { "image": "data/prediction/urban_day_rittenhouse/image/000692.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000692.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.0, "y1": 197.0, "x2": 474.5, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000692" }, { "image": "data/prediction/urban_day_rittenhouse/image/000692.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000692.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.0, "y1": 197.0, "x2": 474.5, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000692" }, { "image": "data/prediction/urban_day_rittenhouse/image/000693.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000693.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 422.0, "y1": 197.5, "x2": 497.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000693" }, { "image": "data/prediction/urban_day_rittenhouse/image/000693.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000693.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 422.0, "y1": 197.5, "x2": 497.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000693" }, { "image": "data/prediction/urban_day_rittenhouse/image/000694.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000694.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 431.0, "y1": 199.0, "x2": 527.0, "y2": 263.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000694" }, { "image": "data/prediction/urban_day_rittenhouse/image/000694.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000694.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.0, "y1": 199.0, "x2": 527.0, "y2": 263.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000694" }, { "image": "data/prediction/urban_day_rittenhouse/image/000695.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000695.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 445.5, "y1": 199.0, "x2": 575.0, "y2": 284.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000695" }, { "image": "data/prediction/urban_day_rittenhouse/image/000695.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000695.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 445.5, "y1": 199.0, "x2": 575.0, "y2": 284.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000695" }, { "image": "data/prediction/urban_day_rittenhouse/image/000696.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000696.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 471.5, "y1": 201.0, "x2": 639.5, "y2": 329.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000696" }, { "image": "data/prediction/urban_day_rittenhouse/image/000696.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000696.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 471.5, "y1": 201.0, "x2": 639.5, "y2": 329.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000696" }, { "image": "data/prediction/urban_day_rittenhouse/image/000697.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000697.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 407.0, "y1": 196.5, "x2": 458.5, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000697" }, { "image": "data/prediction/urban_day_rittenhouse/image/000697.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000697.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.0, "y1": 196.5, "x2": 458.5, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000697" }, { "image": "data/prediction/urban_day_rittenhouse/image/000698.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000698.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 411.0, "y1": 198.5, "x2": 472.0, "y2": 242.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000698" }, { "image": "data/prediction/urban_day_rittenhouse/image/000698.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000698.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.0, "y1": 198.5, "x2": 472.0, "y2": 242.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000698" }, { "image": "data/prediction/urban_day_rittenhouse/image/000699.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000699.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.5, "y1": 201.0, "x2": 499.5, "y2": 254.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000699" }, { "image": "data/prediction/urban_day_rittenhouse/image/000699.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000699.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.5, "y1": 201.0, "x2": 499.5, "y2": 254.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000699" }, { "image": "data/prediction/urban_day_rittenhouse/image/000700.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000700.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.0, "y1": 201.0, "x2": 543.0, "y2": 271.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000700" }, { "image": "data/prediction/urban_day_rittenhouse/image/000700.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000700.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.0, "y1": 201.0, "x2": 543.0, "y2": 271.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000700" }, { "image": "data/prediction/urban_day_rittenhouse/image/000701.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000701.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 470.5, "y1": 205.5, "x2": 639.5, "y2": 312.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000701" }, { "image": "data/prediction/urban_day_rittenhouse/image/000701.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000701.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 470.5, "y1": 205.5, "x2": 639.5, "y2": 312.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000701" }, { "image": "data/prediction/urban_day_rittenhouse/image/000702.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000702.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 40.5, "y1": 184.5, "x2": 122.5, "y2": 218.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000702" }, { "image": "data/prediction/urban_day_rittenhouse/image/000702.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000702.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 40.5, "y1": 184.5, "x2": 122.5, "y2": 218.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000702" }, { "image": "data/prediction/urban_day_rittenhouse/image/000703.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000703.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 175.5, "x2": 76.5, "y2": 214.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000703" }, { "image": "data/prediction/urban_day_rittenhouse/image/000703.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000703.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 175.5, "x2": 76.5, "y2": 214.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000703" }, { "image": "data/prediction/urban_day_rittenhouse/image/000704.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000704.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 577.5, "y1": 210.0, "x2": 639.5, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000704" }, { "image": "data/prediction/urban_day_rittenhouse/image/000704.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000704.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 577.5, "y1": 210.0, "x2": 639.5, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000704" }, { "image": "data/prediction/urban_day_rittenhouse/image/000705.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000705.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 431.0, "y1": 203.5, "x2": 606.0, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000705" }, { "image": "data/prediction/urban_day_rittenhouse/image/000705.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000705.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 431.0, "y1": 203.5, "x2": 606.0, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000705" }, { "image": "data/prediction/urban_day_rittenhouse/image/000706.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000706.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 221.5, "y1": 192.5, "x2": 413.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000706" }, { "image": "data/prediction/urban_day_rittenhouse/image/000706.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000706.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 221.5, "y1": 192.5, "x2": 413.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000706" }, { "image": "data/prediction/urban_day_rittenhouse/image/000707.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000707.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 180.0, "x2": 153.0, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000707" }, { "image": "data/prediction/urban_day_rittenhouse/image/000707.jpg", "event": "data/prediction/urban_day_rittenhouse/event/000707.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 180.0, "x2": 153.0, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000707" }, { "image": "data/prediction/urban_night_rittenhouse/image/000000.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000000.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 582.5, "y1": 208.5, "x2": 639.5, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000708" }, { "image": "data/prediction/urban_night_rittenhouse/image/000000.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000000.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 582.5, "y1": 208.5, "x2": 639.5, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000708" }, { "image": "data/prediction/urban_night_rittenhouse/image/000001.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000001.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 471.0, "y1": 208.0, "x2": 622.0, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000709" }, { "image": "data/prediction/urban_night_rittenhouse/image/000001.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000001.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 471.0, "y1": 208.0, "x2": 622.0, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000709" }, { "image": "data/prediction/urban_night_rittenhouse/image/000002.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000002.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 328.0, "y1": 202.0, "x2": 492.5, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000710" }, { "image": "data/prediction/urban_night_rittenhouse/image/000002.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000002.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 328.0, "y1": 202.0, "x2": 492.5, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000710" }, { "image": "data/prediction/urban_night_rittenhouse/image/000003.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000003.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 161.0, "y1": 194.5, "x2": 324.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000711" }, { "image": "data/prediction/urban_night_rittenhouse/image/000003.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000003.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 161.0, "y1": 194.5, "x2": 324.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000711" }, { "image": "data/prediction/urban_night_rittenhouse/image/000004.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000004.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 3.5, "y1": 185.5, "x2": 153.0, "y2": 240.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000712" }, { "image": "data/prediction/urban_night_rittenhouse/image/000004.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000004.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 3.5, "y1": 185.5, "x2": 153.0, "y2": 240.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000712" }, { "image": "data/prediction/urban_night_rittenhouse/image/000005.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000005.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 595.5, "y1": 206.0, "x2": 639.5, "y2": 266.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000713" }, { "image": "data/prediction/urban_night_rittenhouse/image/000005.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000005.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 595.5, "y1": 206.0, "x2": 639.5, "y2": 266.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000713" }, { "image": "data/prediction/urban_night_rittenhouse/image/000006.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000006.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 499.0, "y1": 202.5, "x2": 639.5, "y2": 266.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000714" }, { "image": "data/prediction/urban_night_rittenhouse/image/000006.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000006.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 499.0, "y1": 202.5, "x2": 639.5, "y2": 266.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000714" }, { "image": "data/prediction/urban_night_rittenhouse/image/000007.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000007.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 305.5, "y1": 195.5, "x2": 479.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000715" }, { "image": "data/prediction/urban_night_rittenhouse/image/000007.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000007.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 305.5, "y1": 195.5, "x2": 479.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000715" }, { "image": "data/prediction/urban_night_rittenhouse/image/000008.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000008.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 130.0, "y1": 185.0, "x2": 301.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000716" }, { "image": "data/prediction/urban_night_rittenhouse/image/000008.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000008.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 130.0, "y1": 185.0, "x2": 301.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000716" }, { "image": "data/prediction/urban_night_rittenhouse/image/000009.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000009.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 175.5, "x2": 140.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000717" }, { "image": "data/prediction/urban_night_rittenhouse/image/000009.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000009.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 175.5, "x2": 140.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000717" }, { "image": "data/prediction/urban_night_rittenhouse/image/000010.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000010.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 564.5, "y1": 211.0, "x2": 639.5, "y2": 271.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000718" }, { "image": "data/prediction/urban_night_rittenhouse/image/000010.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000010.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 564.5, "y1": 211.0, "x2": 639.5, "y2": 271.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000718" }, { "image": "data/prediction/urban_night_rittenhouse/image/000011.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000011.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 369.5, "y1": 203.5, "x2": 542.0, "y2": 266.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000719" }, { "image": "data/prediction/urban_night_rittenhouse/image/000011.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000011.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 369.5, "y1": 203.5, "x2": 542.0, "y2": 266.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000719" }, { "image": "data/prediction/urban_night_rittenhouse/image/000012.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000012.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 135.0, "y1": 194.0, "x2": 304.5, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000720" }, { "image": "data/prediction/urban_night_rittenhouse/image/000012.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000012.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 135.0, "y1": 194.0, "x2": 304.5, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000720" }, { "image": "data/prediction/urban_night_rittenhouse/image/000013.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000013.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 521.5, "y1": 206.0, "x2": 639.5, "y2": 272.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000721" }, { "image": "data/prediction/urban_night_rittenhouse/image/000013.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000013.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 521.5, "y1": 206.0, "x2": 639.5, "y2": 272.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000721" }, { "image": "data/prediction/urban_night_rittenhouse/image/000014.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000014.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 367.5, "y1": 199.0, "x2": 542.5, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000722" }, { "image": "data/prediction/urban_night_rittenhouse/image/000014.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000014.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 367.5, "y1": 199.0, "x2": 542.5, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000722" }, { "image": "data/prediction/urban_night_rittenhouse/image/000015.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000015.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 200.0, "y1": 192.0, "x2": 380.5, "y2": 255.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000723" }, { "image": "data/prediction/urban_night_rittenhouse/image/000015.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000015.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 200.0, "y1": 192.0, "x2": 380.5, "y2": 255.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000723" }, { "image": "data/prediction/urban_night_rittenhouse/image/000016.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000016.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 49.0, "y1": 186.0, "x2": 217.5, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000724" }, { "image": "data/prediction/urban_night_rittenhouse/image/000016.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000016.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 49.0, "y1": 186.0, "x2": 217.5, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000724" }, { "image": "data/prediction/urban_night_rittenhouse/image/000017.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000017.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 99.5, "y1": 199.5, "x2": 198.0, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000725" }, { "image": "data/prediction/urban_night_rittenhouse/image/000017.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000017.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 99.5, "y1": 199.5, "x2": 198.0, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000725" }, { "image": "data/prediction/urban_night_rittenhouse/image/000018.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000018.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 319.0, "y1": 201.5, "x2": 443.0, "y2": 270.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000726" }, { "image": "data/prediction/urban_night_rittenhouse/image/000018.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000018.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 319.0, "y1": 201.5, "x2": 443.0, "y2": 270.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000726" }, { "image": "data/prediction/urban_night_rittenhouse/image/000019.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000019.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 479.0, "y1": 212.0, "x2": 639.5, "y2": 318.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000727" }, { "image": "data/prediction/urban_night_rittenhouse/image/000019.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000019.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 479.0, "y1": 212.0, "x2": 639.5, "y2": 318.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000727" }, { "image": "data/prediction/urban_night_rittenhouse/image/000020.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000020.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 183.0, "y1": 190.5, "x2": 238.0, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000728" }, { "image": "data/prediction/urban_night_rittenhouse/image/000020.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000020.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 183.0, "y1": 190.5, "x2": 238.0, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000728" }, { "image": "data/prediction/urban_night_rittenhouse/image/000021.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000021.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 329.5, "y1": 190.5, "x2": 389.0, "y2": 245.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000729" }, { "image": "data/prediction/urban_night_rittenhouse/image/000021.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000021.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 329.5, "y1": 190.5, "x2": 389.0, "y2": 245.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000729" }, { "image": "data/prediction/urban_night_rittenhouse/image/000022.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000022.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 362.0, "y1": 195.5, "x2": 429.0, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000730" }, { "image": "data/prediction/urban_night_rittenhouse/image/000022.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000022.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 362.0, "y1": 195.5, "x2": 429.0, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000730" }, { "image": "data/prediction/urban_night_rittenhouse/image/000023.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000023.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 386.5, "y1": 197.0, "x2": 463.0, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000731" }, { "image": "data/prediction/urban_night_rittenhouse/image/000023.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000023.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 386.5, "y1": 197.0, "x2": 463.0, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000731" }, { "image": "data/prediction/urban_night_rittenhouse/image/000024.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000024.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 411.5, "y1": 191.5, "x2": 503.0, "y2": 275.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000732" }, { "image": "data/prediction/urban_night_rittenhouse/image/000024.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000024.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.5, "y1": 191.5, "x2": 503.0, "y2": 275.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000732" }, { "image": "data/prediction/urban_night_rittenhouse/image/000025.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000025.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 423.0, "y1": 194.5, "x2": 551.0, "y2": 300.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000733" }, { "image": "data/prediction/urban_night_rittenhouse/image/000025.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000025.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 423.0, "y1": 194.5, "x2": 551.0, "y2": 300.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000733" }, { "image": "data/prediction/urban_night_rittenhouse/image/000026.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000026.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 421.5, "y1": 203.0, "x2": 586.5, "y2": 328.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000734" }, { "image": "data/prediction/urban_night_rittenhouse/image/000026.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000026.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.5, "y1": 203.0, "x2": 586.5, "y2": 328.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000734" }, { "image": "data/prediction/urban_night_rittenhouse/image/000027.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000027.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 413.0, "y1": 204.5, "x2": 602.0, "y2": 342.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000735" }, { "image": "data/prediction/urban_night_rittenhouse/image/000027.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000027.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 413.0, "y1": 204.5, "x2": 602.0, "y2": 342.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000735" }, { "image": "data/prediction/urban_night_rittenhouse/image/000028.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000028.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 392.5, "y1": 206.0, "x2": 609.5, "y2": 354.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000736" }, { "image": "data/prediction/urban_night_rittenhouse/image/000028.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000028.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 392.5, "y1": 206.0, "x2": 609.5, "y2": 354.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000736" }, { "image": "data/prediction/urban_night_rittenhouse/image/000029.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000029.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 365.5, "y1": 200.0, "x2": 598.0, "y2": 357.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000737" }, { "image": "data/prediction/urban_night_rittenhouse/image/000029.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000029.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 365.5, "y1": 200.0, "x2": 598.0, "y2": 357.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000737" }, { "image": "data/prediction/urban_night_rittenhouse/image/000030.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000030.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 339.5, "y1": 204.0, "x2": 569.0, "y2": 369.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000738" }, { "image": "data/prediction/urban_night_rittenhouse/image/000030.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000030.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 339.5, "y1": 204.0, "x2": 569.0, "y2": 369.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000738" }, { "image": "data/prediction/urban_night_rittenhouse/image/000031.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000031.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 323.5, "y1": 211.5, "x2": 548.0, "y2": 390.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000739" }, { "image": "data/prediction/urban_night_rittenhouse/image/000031.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000031.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.5, "y1": 211.5, "x2": 548.0, "y2": 390.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000739" }, { "image": "data/prediction/urban_night_rittenhouse/image/000032.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000032.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 317.0, "y1": 209.5, "x2": 516.5, "y2": 381.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000740" }, { "image": "data/prediction/urban_night_rittenhouse/image/000032.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000032.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 317.0, "y1": 209.5, "x2": 516.5, "y2": 381.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000740" }, { "image": "data/prediction/urban_night_rittenhouse/image/000033.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000033.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 307.5, "y1": 208.5, "x2": 494.0, "y2": 375.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000741" }, { "image": "data/prediction/urban_night_rittenhouse/image/000033.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000033.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.5, "y1": 208.5, "x2": 494.0, "y2": 375.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000741" }, { "image": "data/prediction/urban_night_rittenhouse/image/000034.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000034.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 297.0, "y1": 207.5, "x2": 476.0, "y2": 374.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000742" }, { "image": "data/prediction/urban_night_rittenhouse/image/000034.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000034.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 297.0, "y1": 207.5, "x2": 476.0, "y2": 374.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000742" }, { "image": "data/prediction/urban_night_rittenhouse/image/000035.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000035.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 281.5, "y1": 203.0, "x2": 451.5, "y2": 361.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000743" }, { "image": "data/prediction/urban_night_rittenhouse/image/000035.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000035.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 281.5, "y1": 203.0, "x2": 451.5, "y2": 361.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000743" }, { "image": "data/prediction/urban_night_rittenhouse/image/000036.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000036.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 277.0, "y1": 205.0, "x2": 428.5, "y2": 344.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000744" }, { "image": "data/prediction/urban_night_rittenhouse/image/000036.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000036.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 277.0, "y1": 205.0, "x2": 428.5, "y2": 344.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000744" }, { "image": "data/prediction/urban_night_rittenhouse/image/000037.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000037.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 376.5, "y1": 193.0, "x2": 463.0, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000745" }, { "image": "data/prediction/urban_night_rittenhouse/image/000037.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000037.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 376.5, "y1": 193.0, "x2": 463.0, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000745" }, { "image": "data/prediction/urban_night_rittenhouse/image/000038.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000038.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 389.5, "y1": 197.5, "x2": 505.5, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000746" }, { "image": "data/prediction/urban_night_rittenhouse/image/000038.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000038.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 389.5, "y1": 197.5, "x2": 505.5, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000746" }, { "image": "data/prediction/urban_night_rittenhouse/image/000039.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000039.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 411.0, "y1": 199.5, "x2": 531.5, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000747" }, { "image": "data/prediction/urban_night_rittenhouse/image/000039.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000039.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.0, "y1": 199.5, "x2": 531.5, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000747" }, { "image": "data/prediction/urban_night_rittenhouse/image/000040.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000040.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 359.5, "y1": 190.5, "x2": 468.5, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000748" }, { "image": "data/prediction/urban_night_rittenhouse/image/000040.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000040.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 359.5, "y1": 190.5, "x2": 468.5, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000748" }, { "image": "data/prediction/urban_night_rittenhouse/image/000041.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000041.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.0, "y1": 195.0, "x2": 544.0, "y2": 268.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000749" }, { "image": "data/prediction/urban_night_rittenhouse/image/000041.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000041.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.0, "y1": 195.0, "x2": 544.0, "y2": 268.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000749" }, { "image": "data/prediction/urban_night_rittenhouse/image/000042.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000042.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.0, "y1": 193.0, "x2": 602.0, "y2": 297.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000750" }, { "image": "data/prediction/urban_night_rittenhouse/image/000042.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000042.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.0, "y1": 193.0, "x2": 602.0, "y2": 297.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000750" }, { "image": "data/prediction/urban_night_rittenhouse/image/000043.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000043.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 285.5, "y1": 203.0, "x2": 428.0, "y2": 335.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000751" }, { "image": "data/prediction/urban_night_rittenhouse/image/000043.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000043.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 285.5, "y1": 203.0, "x2": 428.0, "y2": 335.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000751" }, { "image": "data/prediction/urban_night_rittenhouse/image/000044.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000044.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 283.0, "y1": 200.5, "x2": 422.5, "y2": 328.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000752" }, { "image": "data/prediction/urban_night_rittenhouse/image/000044.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000044.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 283.0, "y1": 200.5, "x2": 422.5, "y2": 328.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000752" }, { "image": "data/prediction/urban_night_rittenhouse/image/000045.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000045.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 284.0, "y1": 198.5, "x2": 417.0, "y2": 320.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000753" }, { "image": "data/prediction/urban_night_rittenhouse/image/000045.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000045.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 284.0, "y1": 198.5, "x2": 417.0, "y2": 320.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000753" }, { "image": "data/prediction/urban_night_rittenhouse/image/000046.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000046.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 284.5, "y1": 199.0, "x2": 411.0, "y2": 315.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000754" }, { "image": "data/prediction/urban_night_rittenhouse/image/000046.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000046.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 284.5, "y1": 199.0, "x2": 411.0, "y2": 315.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000754" }, { "image": "data/prediction/urban_night_rittenhouse/image/000047.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000047.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 273.5, "y1": 199.0, "x2": 400.0, "y2": 314.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000755" }, { "image": "data/prediction/urban_night_rittenhouse/image/000047.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000047.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 273.5, "y1": 199.0, "x2": 400.0, "y2": 314.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000755" }, { "image": "data/prediction/urban_night_rittenhouse/image/000048.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000048.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 273.0, "y1": 197.5, "x2": 399.0, "y2": 312.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000756" }, { "image": "data/prediction/urban_night_rittenhouse/image/000048.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000048.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 273.0, "y1": 197.5, "x2": 399.0, "y2": 312.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000756" }, { "image": "data/prediction/urban_night_rittenhouse/image/000049.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000049.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 274.0, "y1": 197.5, "x2": 392.0, "y2": 304.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000757" }, { "image": "data/prediction/urban_night_rittenhouse/image/000049.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000049.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 274.0, "y1": 197.5, "x2": 392.0, "y2": 304.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000757" }, { "image": "data/prediction/urban_night_rittenhouse/image/000050.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000050.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 280.5, "y1": 201.5, "x2": 390.5, "y2": 302.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000758" }, { "image": "data/prediction/urban_night_rittenhouse/image/000050.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000050.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.5, "y1": 201.5, "x2": 390.5, "y2": 302.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000758" }, { "image": "data/prediction/urban_night_rittenhouse/image/000051.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000051.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 288.5, "y1": 200.5, "x2": 390.5, "y2": 293.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000759" }, { "image": "data/prediction/urban_night_rittenhouse/image/000051.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000051.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 288.5, "y1": 200.5, "x2": 390.5, "y2": 293.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000759" }, { "image": "data/prediction/urban_night_rittenhouse/image/000052.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000052.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 295.5, "y1": 193.0, "x2": 390.0, "y2": 279.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000760" }, { "image": "data/prediction/urban_night_rittenhouse/image/000052.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000052.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 295.5, "y1": 193.0, "x2": 390.0, "y2": 279.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000760" }, { "image": "data/prediction/urban_night_rittenhouse/image/000053.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000053.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 461.0, "y1": 212.0, "x2": 625.0, "y2": 324.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000761" }, { "image": "data/prediction/urban_night_rittenhouse/image/000053.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000053.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 461.0, "y1": 212.0, "x2": 625.0, "y2": 324.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000761" }, { "image": "data/prediction/urban_night_rittenhouse/image/000054.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000054.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 298.5, "y1": 194.0, "x2": 390.0, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000762" }, { "image": "data/prediction/urban_night_rittenhouse/image/000054.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000054.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 298.5, "y1": 194.0, "x2": 390.0, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000762" }, { "image": "data/prediction/urban_night_rittenhouse/image/000055.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000055.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 293.5, "y1": 198.0, "x2": 390.0, "y2": 285.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000763" }, { "image": "data/prediction/urban_night_rittenhouse/image/000055.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000055.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 293.5, "y1": 198.0, "x2": 390.0, "y2": 285.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000763" }, { "image": "data/prediction/urban_night_rittenhouse/image/000056.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000056.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 305.0, "y1": 198.0, "x2": 410.0, "y2": 295.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000764" }, { "image": "data/prediction/urban_night_rittenhouse/image/000056.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000056.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.0, "y1": 198.0, "x2": 410.0, "y2": 295.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000764" }, { "image": "data/prediction/urban_night_rittenhouse/image/000057.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000057.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 299.0, "y1": 197.5, "x2": 408.5, "y2": 298.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000765" }, { "image": "data/prediction/urban_night_rittenhouse/image/000057.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000057.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.0, "y1": 197.5, "x2": 408.5, "y2": 298.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000765" }, { "image": "data/prediction/urban_night_rittenhouse/image/000058.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000058.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 302.0, "y1": 198.0, "x2": 409.5, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000766" }, { "image": "data/prediction/urban_night_rittenhouse/image/000058.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000058.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 302.0, "y1": 198.0, "x2": 409.5, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000766" }, { "image": "data/prediction/urban_night_rittenhouse/image/000059.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000059.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 306.5, "y1": 198.5, "x2": 410.5, "y2": 294.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000767" }, { "image": "data/prediction/urban_night_rittenhouse/image/000059.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000059.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.5, "y1": 198.5, "x2": 410.5, "y2": 294.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000767" }, { "image": "data/prediction/urban_night_rittenhouse/image/000060.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000060.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 304.0, "y1": 198.5, "x2": 408.5, "y2": 293.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000768" }, { "image": "data/prediction/urban_night_rittenhouse/image/000060.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000060.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.0, "y1": 198.5, "x2": 408.5, "y2": 293.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000768" }, { "image": "data/prediction/urban_night_rittenhouse/image/000061.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000061.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 306.5, "y1": 198.0, "x2": 413.5, "y2": 296.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000769" }, { "image": "data/prediction/urban_night_rittenhouse/image/000061.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000061.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.5, "y1": 198.0, "x2": 413.5, "y2": 296.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000769" }, { "image": "data/prediction/urban_night_rittenhouse/image/000062.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000062.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 304.5, "y1": 197.0, "x2": 414.5, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000770" }, { "image": "data/prediction/urban_night_rittenhouse/image/000062.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000062.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.5, "y1": 197.0, "x2": 414.5, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000770" }, { "image": "data/prediction/urban_night_rittenhouse/image/000063.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000063.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 306.5, "y1": 199.0, "x2": 413.5, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000771" }, { "image": "data/prediction/urban_night_rittenhouse/image/000063.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000063.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.5, "y1": 199.0, "x2": 413.5, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000771" }, { "image": "data/prediction/urban_night_rittenhouse/image/000064.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000064.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 398.0, "y1": 191.0, "x2": 456.5, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000772" }, { "image": "data/prediction/urban_night_rittenhouse/image/000064.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000064.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 398.0, "y1": 191.0, "x2": 456.5, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000772" }, { "image": "data/prediction/urban_night_rittenhouse/image/000065.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000065.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.5, "y1": 192.5, "x2": 463.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000773" }, { "image": "data/prediction/urban_night_rittenhouse/image/000065.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000065.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.5, "y1": 192.5, "x2": 463.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000773" }, { "image": "data/prediction/urban_night_rittenhouse/image/000066.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000066.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 423.0, "y1": 195.5, "x2": 492.5, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000774" }, { "image": "data/prediction/urban_night_rittenhouse/image/000066.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000066.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 423.0, "y1": 195.5, "x2": 492.5, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000774" }, { "image": "data/prediction/urban_night_rittenhouse/image/000067.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000067.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.5, "y1": 195.0, "x2": 509.0, "y2": 261.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000775" }, { "image": "data/prediction/urban_night_rittenhouse/image/000067.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000067.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.5, "y1": 195.0, "x2": 509.0, "y2": 261.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000775" }, { "image": "data/prediction/urban_night_rittenhouse/image/000068.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000068.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.5, "y1": 196.5, "x2": 528.0, "y2": 271.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000776" }, { "image": "data/prediction/urban_night_rittenhouse/image/000068.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000068.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.5, "y1": 196.5, "x2": 528.0, "y2": 271.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000776" }, { "image": "data/prediction/urban_night_rittenhouse/image/000069.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000069.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 410.0, "y1": 196.5, "x2": 537.5, "y2": 276.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000777" }, { "image": "data/prediction/urban_night_rittenhouse/image/000069.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000069.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 410.0, "y1": 196.5, "x2": 537.5, "y2": 276.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000777" }, { "image": "data/prediction/urban_night_rittenhouse/image/000070.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000070.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 406.5, "y1": 197.0, "x2": 548.0, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000778" }, { "image": "data/prediction/urban_night_rittenhouse/image/000070.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000070.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 406.5, "y1": 197.0, "x2": 548.0, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000778" }, { "image": "data/prediction/urban_night_rittenhouse/image/000071.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000071.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 411.5, "y1": 197.0, "x2": 562.5, "y2": 290.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000779" }, { "image": "data/prediction/urban_night_rittenhouse/image/000071.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000071.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.5, "y1": 197.0, "x2": 562.5, "y2": 290.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000779" }, { "image": "data/prediction/urban_night_rittenhouse/image/000072.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000072.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 411.5, "y1": 197.0, "x2": 571.5, "y2": 294.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000780" }, { "image": "data/prediction/urban_night_rittenhouse/image/000072.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000072.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.5, "y1": 197.0, "x2": 571.5, "y2": 294.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000780" }, { "image": "data/prediction/urban_night_rittenhouse/image/000073.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000073.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 192.0, "y1": 184.0, "x2": 246.5, "y2": 204.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000781" }, { "image": "data/prediction/urban_night_rittenhouse/image/000073.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000073.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 192.0, "y1": 184.0, "x2": 246.5, "y2": 204.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000781" }, { "image": "data/prediction/urban_night_rittenhouse/image/000074.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000074.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 531.5, "y1": 204.5, "x2": 639.5, "y2": 311.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000782" }, { "image": "data/prediction/urban_night_rittenhouse/image/000074.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000074.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 531.5, "y1": 204.5, "x2": 639.5, "y2": 311.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000782" }, { "image": "data/prediction/urban_night_rittenhouse/image/000075.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000075.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 92.0, "y1": 185.0, "x2": 163.0, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000783" }, { "image": "data/prediction/urban_night_rittenhouse/image/000075.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000075.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 92.0, "y1": 185.0, "x2": 163.0, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000783" }, { "image": "data/prediction/urban_night_rittenhouse/image/000076.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000076.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 87.5, "y1": 184.5, "x2": 160.0, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000784" }, { "image": "data/prediction/urban_night_rittenhouse/image/000076.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000076.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 87.5, "y1": 184.5, "x2": 160.0, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000784" }, { "image": "data/prediction/urban_night_rittenhouse/image/000077.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000077.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 84.0, "y1": 183.0, "x2": 159.0, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000785" }, { "image": "data/prediction/urban_night_rittenhouse/image/000077.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000077.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 84.0, "y1": 183.0, "x2": 159.0, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000785" }, { "image": "data/prediction/urban_night_rittenhouse/image/000078.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000078.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 78.0, "y1": 182.5, "x2": 154.0, "y2": 210.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000786" }, { "image": "data/prediction/urban_night_rittenhouse/image/000078.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000078.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 78.0, "y1": 182.5, "x2": 154.0, "y2": 210.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000786" }, { "image": "data/prediction/urban_night_rittenhouse/image/000079.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000079.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 79.0, "y1": 183.5, "x2": 154.0, "y2": 211.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000787" }, { "image": "data/prediction/urban_night_rittenhouse/image/000079.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000079.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 79.0, "y1": 183.5, "x2": 154.0, "y2": 211.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000787" }, { "image": "data/prediction/urban_night_rittenhouse/image/000080.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000080.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 84.5, "y1": 187.5, "x2": 159.5, "y2": 214.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000788" }, { "image": "data/prediction/urban_night_rittenhouse/image/000080.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000080.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 84.5, "y1": 187.5, "x2": 159.5, "y2": 214.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000788" }, { "image": "data/prediction/urban_night_rittenhouse/image/000081.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000081.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 79.0, "y1": 184.5, "x2": 154.0, "y2": 212.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000789" }, { "image": "data/prediction/urban_night_rittenhouse/image/000081.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000081.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 79.0, "y1": 184.5, "x2": 154.0, "y2": 212.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000789" }, { "image": "data/prediction/urban_night_rittenhouse/image/000082.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000082.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 64.5, "y1": 187.0, "x2": 140.0, "y2": 214.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000790" }, { "image": "data/prediction/urban_night_rittenhouse/image/000082.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000082.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 64.5, "y1": 187.0, "x2": 140.0, "y2": 214.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000790" }, { "image": "data/prediction/urban_night_rittenhouse/image/000083.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000083.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 53.5, "y1": 185.0, "x2": 133.5, "y2": 214.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000791" }, { "image": "data/prediction/urban_night_rittenhouse/image/000083.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000083.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 53.5, "y1": 185.0, "x2": 133.5, "y2": 214.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000791" }, { "image": "data/prediction/urban_night_rittenhouse/image/000084.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000084.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 304.5, "y1": 197.5, "x2": 410.0, "y2": 295.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000792" }, { "image": "data/prediction/urban_night_rittenhouse/image/000084.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000084.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 304.5, "y1": 197.5, "x2": 410.0, "y2": 295.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000792" }, { "image": "data/prediction/urban_night_rittenhouse/image/000085.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000085.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 305.0, "y1": 199.0, "x2": 404.0, "y2": 288.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000793" }, { "image": "data/prediction/urban_night_rittenhouse/image/000085.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000085.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.0, "y1": 199.0, "x2": 404.0, "y2": 288.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000793" }, { "image": "data/prediction/urban_night_rittenhouse/image/000086.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000086.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 306.0, "y1": 200.0, "x2": 394.0, "y2": 279.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000794" }, { "image": "data/prediction/urban_night_rittenhouse/image/000086.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000086.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.0, "y1": 200.0, "x2": 394.0, "y2": 279.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000794" }, { "image": "data/prediction/urban_night_rittenhouse/image/000087.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000087.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 309.0, "y1": 199.0, "x2": 389.5, "y2": 271.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000795" }, { "image": "data/prediction/urban_night_rittenhouse/image/000087.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000087.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.0, "y1": 199.0, "x2": 389.5, "y2": 271.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000795" }, { "image": "data/prediction/urban_night_rittenhouse/image/000088.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000088.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 309.0, "y1": 199.0, "x2": 385.0, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000796" }, { "image": "data/prediction/urban_night_rittenhouse/image/000088.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000088.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.0, "y1": 199.0, "x2": 385.0, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000796" }, { "image": "data/prediction/urban_night_rittenhouse/image/000089.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000089.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 312.5, "y1": 199.5, "x2": 387.0, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000797" }, { "image": "data/prediction/urban_night_rittenhouse/image/000089.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000089.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 312.5, "y1": 199.5, "x2": 387.0, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000797" }, { "image": "data/prediction/urban_night_rittenhouse/image/000090.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000090.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 313.0, "y1": 196.5, "x2": 385.5, "y2": 263.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000798" }, { "image": "data/prediction/urban_night_rittenhouse/image/000090.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000090.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 313.0, "y1": 196.5, "x2": 385.5, "y2": 263.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000798" }, { "image": "data/prediction/urban_night_rittenhouse/image/000091.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000091.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 313.5, "y1": 195.0, "x2": 383.5, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000799" }, { "image": "data/prediction/urban_night_rittenhouse/image/000091.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000091.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 313.5, "y1": 195.0, "x2": 383.5, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000799" }, { "image": "data/prediction/urban_night_rittenhouse/image/000092.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000092.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 310.5, "y1": 195.5, "x2": 379.5, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000800" }, { "image": "data/prediction/urban_night_rittenhouse/image/000092.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000092.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 310.5, "y1": 195.5, "x2": 379.5, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000800" }, { "image": "data/prediction/urban_night_rittenhouse/image/000093.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000093.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 310.5, "y1": 196.5, "x2": 380.0, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000801" }, { "image": "data/prediction/urban_night_rittenhouse/image/000093.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000093.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 310.5, "y1": 196.5, "x2": 380.0, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000801" }, { "image": "data/prediction/urban_night_rittenhouse/image/000094.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000094.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 611.0, "y1": 206.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000802" }, { "image": "data/prediction/urban_night_rittenhouse/image/000094.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000094.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 611.0, "y1": 206.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000802" }, { "image": "data/prediction/urban_night_rittenhouse/image/000095.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000095.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 600.0, "y1": 206.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000803" }, { "image": "data/prediction/urban_night_rittenhouse/image/000095.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000095.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 600.0, "y1": 206.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000803" }, { "image": "data/prediction/urban_night_rittenhouse/image/000096.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000096.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 611.0, "y1": 204.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000804" }, { "image": "data/prediction/urban_night_rittenhouse/image/000096.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000096.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 611.0, "y1": 204.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000804" }, { "image": "data/prediction/urban_night_rittenhouse/image/000097.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000097.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 410.5, "y1": 197.5, "x2": 570.0, "y2": 296.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000805" }, { "image": "data/prediction/urban_night_rittenhouse/image/000097.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000097.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 410.5, "y1": 197.5, "x2": 570.0, "y2": 296.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000805" }, { "image": "data/prediction/urban_night_rittenhouse/image/000098.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000098.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.0, "y1": 197.5, "x2": 562.5, "y2": 291.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000806" }, { "image": "data/prediction/urban_night_rittenhouse/image/000098.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000098.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.0, "y1": 197.5, "x2": 562.5, "y2": 291.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000806" }, { "image": "data/prediction/urban_night_rittenhouse/image/000099.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000099.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 405.5, "y1": 198.5, "x2": 558.0, "y2": 291.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000807" }, { "image": "data/prediction/urban_night_rittenhouse/image/000099.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000099.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 405.5, "y1": 198.5, "x2": 558.0, "y2": 291.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000807" }, { "image": "data/prediction/urban_night_rittenhouse/image/000100.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000100.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 404.0, "y1": 199.5, "x2": 550.5, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000808" }, { "image": "data/prediction/urban_night_rittenhouse/image/000100.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000100.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 404.0, "y1": 199.5, "x2": 550.5, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000808" }, { "image": "data/prediction/urban_night_rittenhouse/image/000101.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000101.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 403.0, "y1": 199.5, "x2": 543.0, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000809" }, { "image": "data/prediction/urban_night_rittenhouse/image/000101.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000101.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.0, "y1": 199.5, "x2": 543.0, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000809" }, { "image": "data/prediction/urban_night_rittenhouse/image/000102.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000102.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 403.5, "y1": 198.0, "x2": 539.5, "y2": 293.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000810" }, { "image": "data/prediction/urban_night_rittenhouse/image/000102.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000102.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.5, "y1": 198.0, "x2": 539.5, "y2": 293.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000810" }, { "image": "data/prediction/urban_night_rittenhouse/image/000103.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000103.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 403.5, "y1": 199.0, "x2": 542.0, "y2": 299.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000811" }, { "image": "data/prediction/urban_night_rittenhouse/image/000103.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000103.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.5, "y1": 199.0, "x2": 542.0, "y2": 299.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000811" }, { "image": "data/prediction/urban_night_rittenhouse/image/000104.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000104.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 405.0, "y1": 197.5, "x2": 530.0, "y2": 295.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000812" }, { "image": "data/prediction/urban_night_rittenhouse/image/000104.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000104.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 405.0, "y1": 197.5, "x2": 530.0, "y2": 295.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000812" }, { "image": "data/prediction/urban_night_rittenhouse/image/000105.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000105.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 403.0, "y1": 201.5, "x2": 524.0, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000813" }, { "image": "data/prediction/urban_night_rittenhouse/image/000105.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000105.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.0, "y1": 201.5, "x2": 524.0, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000813" }, { "image": "data/prediction/urban_night_rittenhouse/image/000106.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000106.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 400.0, "y1": 200.5, "x2": 528.5, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000814" }, { "image": "data/prediction/urban_night_rittenhouse/image/000106.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000106.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 400.0, "y1": 200.5, "x2": 528.5, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000814" }, { "image": "data/prediction/urban_night_rittenhouse/image/000107.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000107.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 261.5, "y1": 195.0, "x2": 327.0, "y2": 233.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000815" }, { "image": "data/prediction/urban_night_rittenhouse/image/000107.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000107.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 261.5, "y1": 195.0, "x2": 327.0, "y2": 233.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000815" }, { "image": "data/prediction/urban_night_rittenhouse/image/000108.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000108.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 212.0, "y1": 193.5, "x2": 292.0, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000816" }, { "image": "data/prediction/urban_night_rittenhouse/image/000108.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000108.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 212.0, "y1": 193.5, "x2": 292.0, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000816" }, { "image": "data/prediction/urban_night_rittenhouse/image/000109.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000109.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 149.5, "y1": 189.5, "x2": 239.5, "y2": 227.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000817" }, { "image": "data/prediction/urban_night_rittenhouse/image/000109.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000109.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 149.5, "y1": 189.5, "x2": 239.5, "y2": 227.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000817" }, { "image": "data/prediction/urban_night_rittenhouse/image/000110.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000110.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 80.0, "y1": 187.0, "x2": 168.0, "y2": 224.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000818" }, { "image": "data/prediction/urban_night_rittenhouse/image/000110.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000110.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 80.0, "y1": 187.0, "x2": 168.0, "y2": 224.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000818" }, { "image": "data/prediction/urban_night_rittenhouse/image/000111.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000111.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 472.5, "y1": 209.0, "x2": 597.0, "y2": 284.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000819" }, { "image": "data/prediction/urban_night_rittenhouse/image/000111.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000111.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 472.5, "y1": 209.0, "x2": 597.0, "y2": 284.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000819" }, { "image": "data/prediction/urban_night_rittenhouse/image/000112.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000112.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 490.5, "y1": 210.0, "x2": 639.5, "y2": 302.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000820" }, { "image": "data/prediction/urban_night_rittenhouse/image/000112.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000112.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 490.5, "y1": 210.0, "x2": 639.5, "y2": 302.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000820" }, { "image": "data/prediction/urban_night_rittenhouse/image/000113.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000113.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 557.5, "y1": 215.5, "x2": 639.5, "y2": 333.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000821" }, { "image": "data/prediction/urban_night_rittenhouse/image/000113.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000113.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 557.5, "y1": 215.5, "x2": 639.5, "y2": 333.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000821" }, { "image": "data/prediction/urban_night_rittenhouse/image/000114.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000114.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 184.0, "x2": 86.0, "y2": 218.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000822" }, { "image": "data/prediction/urban_night_rittenhouse/image/000114.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000114.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 184.0, "x2": 86.0, "y2": 218.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000822" }, { "image": "data/prediction/urban_night_rittenhouse/image/000115.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000115.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 181.0, "x2": 74.0, "y2": 220.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000823" }, { "image": "data/prediction/urban_night_rittenhouse/image/000115.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000115.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 181.0, "x2": 74.0, "y2": 220.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000823" }, { "image": "data/prediction/urban_night_rittenhouse/image/000116.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000116.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 307.0, "y1": 191.5, "x2": 377.0, "y2": 254.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000824" }, { "image": "data/prediction/urban_night_rittenhouse/image/000116.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000116.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.0, "y1": 191.5, "x2": 377.0, "y2": 254.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000824" }, { "image": "data/prediction/urban_night_rittenhouse/image/000117.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000117.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 333.0, "y1": 201.5, "x2": 405.5, "y2": 268.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000825" }, { "image": "data/prediction/urban_night_rittenhouse/image/000117.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000117.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 333.0, "y1": 201.5, "x2": 405.5, "y2": 268.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000825" }, { "image": "data/prediction/urban_night_rittenhouse/image/000118.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000118.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 399.5, "y1": 195.5, "x2": 542.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000826" }, { "image": "data/prediction/urban_night_rittenhouse/image/000118.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000118.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 399.5, "y1": 195.5, "x2": 542.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000826" }, { "image": "data/prediction/urban_night_rittenhouse/image/000119.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000119.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 427.0, "y1": 199.5, "x2": 598.0, "y2": 327.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000827" }, { "image": "data/prediction/urban_night_rittenhouse/image/000119.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000119.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 427.0, "y1": 199.5, "x2": 598.0, "y2": 327.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000827" }, { "image": "data/prediction/urban_night_rittenhouse/image/000120.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000120.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 18.0, "y1": 201.5, "x2": 79.5, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000828" }, { "image": "data/prediction/urban_night_rittenhouse/image/000120.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000120.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 18.0, "y1": 201.5, "x2": 79.5, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000828" }, { "image": "data/prediction/urban_night_rittenhouse/image/000121.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000121.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 107.0, "y1": 208.0, "x2": 197.5, "y2": 287.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000829" }, { "image": "data/prediction/urban_night_rittenhouse/image/000121.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000121.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 107.0, "y1": 208.0, "x2": 197.5, "y2": 287.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000829" }, { "image": "data/prediction/urban_night_rittenhouse/image/000122.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000122.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 44.0, "y1": 217.0, "x2": 230.0, "y2": 355.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000830" }, { "image": "data/prediction/urban_night_rittenhouse/image/000122.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000122.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 44.0, "y1": 217.0, "x2": 230.0, "y2": 355.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000830" }, { "image": "data/prediction/urban_night_rittenhouse/image/000123.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000123.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 204.5, "x2": 37.0, "y2": 289.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000831" }, { "image": "data/prediction/urban_night_rittenhouse/image/000123.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000123.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 204.5, "x2": 37.0, "y2": 289.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000831" }, { "image": "data/prediction/urban_night_rittenhouse/image/000124.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000124.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 217.5, "x2": 175.0, "y2": 348.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000832" }, { "image": "data/prediction/urban_night_rittenhouse/image/000124.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000124.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 217.5, "x2": 175.0, "y2": 348.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000832" }, { "image": "data/prediction/urban_night_rittenhouse/image/000125.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000125.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 220.5, "x2": 40.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000833" }, { "image": "data/prediction/urban_night_rittenhouse/image/000125.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000125.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 220.5, "x2": 40.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000833" }, { "image": "data/prediction/urban_night_rittenhouse/image/000126.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000126.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 176.5, "y1": 197.0, "x2": 243.0, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000834" }, { "image": "data/prediction/urban_night_rittenhouse/image/000126.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000126.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 176.5, "y1": 197.0, "x2": 243.0, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000834" }, { "image": "data/prediction/urban_night_rittenhouse/image/000127.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000127.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 178.0, "y1": 196.0, "x2": 267.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000835" }, { "image": "data/prediction/urban_night_rittenhouse/image/000127.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000127.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 178.0, "y1": 196.0, "x2": 267.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000835" }, { "image": "data/prediction/urban_night_rittenhouse/image/000128.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000128.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 94.5, "y1": 202.0, "x2": 244.0, "y2": 313.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000836" }, { "image": "data/prediction/urban_night_rittenhouse/image/000128.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000128.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 94.5, "y1": 202.0, "x2": 244.0, "y2": 313.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000836" }, { "image": "data/prediction/urban_night_rittenhouse/image/000129.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000129.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 212.0, "y1": 195.0, "x2": 277.0, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000837" }, { "image": "data/prediction/urban_night_rittenhouse/image/000129.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000129.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 212.0, "y1": 195.0, "x2": 277.0, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000837" }, { "image": "data/prediction/urban_night_rittenhouse/image/000130.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000130.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 169.5, "y1": 199.5, "x2": 263.0, "y2": 274.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000838" }, { "image": "data/prediction/urban_night_rittenhouse/image/000130.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000130.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 169.5, "y1": 199.5, "x2": 263.0, "y2": 274.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000838" }, { "image": "data/prediction/urban_night_rittenhouse/image/000131.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000131.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 244.5, "y1": 195.0, "x2": 301.0, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000839" }, { "image": "data/prediction/urban_night_rittenhouse/image/000131.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000131.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 244.5, "y1": 195.0, "x2": 301.0, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000839" }, { "image": "data/prediction/urban_night_rittenhouse/image/000132.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000132.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 212.5, "y1": 197.5, "x2": 291.5, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000840" }, { "image": "data/prediction/urban_night_rittenhouse/image/000132.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000132.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 212.5, "y1": 197.5, "x2": 291.5, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000840" }, { "image": "data/prediction/urban_night_rittenhouse/image/000133.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000133.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 123.0, "y1": 198.0, "x2": 199.5, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000841" }, { "image": "data/prediction/urban_night_rittenhouse/image/000133.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000133.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 123.0, "y1": 198.0, "x2": 199.5, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000841" }, { "image": "data/prediction/urban_night_rittenhouse/image/000134.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000134.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 132.5, "y1": 202.0, "x2": 243.5, "y2": 284.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000842" }, { "image": "data/prediction/urban_night_rittenhouse/image/000134.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000134.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 132.5, "y1": 202.0, "x2": 243.5, "y2": 284.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000842" }, { "image": "data/prediction/urban_night_rittenhouse/image/000135.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000135.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 208.5, "x2": 208.5, "y2": 369.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000843" }, { "image": "data/prediction/urban_night_rittenhouse/image/000135.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000135.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 208.5, "x2": 208.5, "y2": 369.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000843" }, { "image": "data/prediction/urban_night_rittenhouse/image/000136.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000136.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 217.5, "y1": 192.5, "x2": 286.0, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000844" }, { "image": "data/prediction/urban_night_rittenhouse/image/000136.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000136.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 217.5, "y1": 192.5, "x2": 286.0, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000844" }, { "image": "data/prediction/urban_night_rittenhouse/image/000137.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000137.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 171.5, "y1": 197.5, "x2": 273.0, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000845" }, { "image": "data/prediction/urban_night_rittenhouse/image/000137.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000137.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 171.5, "y1": 197.5, "x2": 273.0, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000845" }, { "image": "data/prediction/urban_night_rittenhouse/image/000138.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000138.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 43.5, "y1": 203.5, "x2": 236.5, "y2": 346.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000846" }, { "image": "data/prediction/urban_night_rittenhouse/image/000138.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000138.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 43.5, "y1": 203.5, "x2": 236.5, "y2": 346.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000846" }, { "image": "data/prediction/urban_night_rittenhouse/image/000139.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000139.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 81.5, "y1": 206.5, "x2": 249.0, "y2": 332.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000847" }, { "image": "data/prediction/urban_night_rittenhouse/image/000139.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000139.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 81.5, "y1": 206.5, "x2": 249.0, "y2": 332.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000847" }, { "image": "data/prediction/urban_night_rittenhouse/image/000140.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000140.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 167.5, "y1": 202.5, "x2": 267.5, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000848" }, { "image": "data/prediction/urban_night_rittenhouse/image/000140.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000140.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 167.5, "y1": 202.5, "x2": 267.5, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000848" }, { "image": "data/prediction/urban_night_rittenhouse/image/000141.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000141.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 38.5, "y1": 213.5, "x2": 237.0, "y2": 358.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000849" }, { "image": "data/prediction/urban_night_rittenhouse/image/000141.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000141.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 38.5, "y1": 213.5, "x2": 237.0, "y2": 358.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000849" }, { "image": "data/prediction/urban_night_rittenhouse/image/000142.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000142.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 230.0, "y1": 197.5, "x2": 305.0, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000850" }, { "image": "data/prediction/urban_night_rittenhouse/image/000142.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000142.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 230.0, "y1": 197.5, "x2": 305.0, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000850" }, { "image": "data/prediction/urban_night_rittenhouse/image/000143.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000143.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 182.5, "y1": 200.5, "x2": 280.0, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000851" }, { "image": "data/prediction/urban_night_rittenhouse/image/000143.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000143.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 182.5, "y1": 200.5, "x2": 280.0, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000851" }, { "image": "data/prediction/urban_night_rittenhouse/image/000144.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000144.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 51.5, "y1": 206.5, "x2": 244.5, "y2": 357.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000852" }, { "image": "data/prediction/urban_night_rittenhouse/image/000144.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000144.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 51.5, "y1": 206.5, "x2": 244.5, "y2": 357.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000852" }, { "image": "data/prediction/urban_night_rittenhouse/image/000145.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000145.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 224.5, "y1": 199.5, "x2": 284.0, "y2": 251.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000853" }, { "image": "data/prediction/urban_night_rittenhouse/image/000145.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000145.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 224.5, "y1": 199.5, "x2": 284.0, "y2": 251.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000853" }, { "image": "data/prediction/urban_night_rittenhouse/image/000146.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000146.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 174.0, "y1": 204.0, "x2": 272.0, "y2": 278.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000854" }, { "image": "data/prediction/urban_night_rittenhouse/image/000146.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000146.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 174.0, "y1": 204.0, "x2": 272.0, "y2": 278.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000854" }, { "image": "data/prediction/urban_night_rittenhouse/image/000147.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000147.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 47.5, "y1": 216.0, "x2": 239.5, "y2": 353.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000855" }, { "image": "data/prediction/urban_night_rittenhouse/image/000147.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000147.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 47.5, "y1": 216.0, "x2": 239.5, "y2": 353.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000855" }, { "image": "data/prediction/urban_night_rittenhouse/image/000148.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000148.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 246.5, "y1": 192.0, "x2": 297.5, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000856" }, { "image": "data/prediction/urban_night_rittenhouse/image/000148.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000148.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 246.5, "y1": 192.0, "x2": 297.5, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000856" }, { "image": "data/prediction/urban_night_rittenhouse/image/000149.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000149.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 207.5, "y1": 192.0, "x2": 286.0, "y2": 255.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000857" }, { "image": "data/prediction/urban_night_rittenhouse/image/000149.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000149.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 207.5, "y1": 192.0, "x2": 286.0, "y2": 255.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000857" }, { "image": "data/prediction/urban_night_rittenhouse/image/000150.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000150.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 130.0, "y1": 195.0, "x2": 263.5, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000858" }, { "image": "data/prediction/urban_night_rittenhouse/image/000150.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000150.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 130.0, "y1": 195.0, "x2": 263.5, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000858" }, { "image": "data/prediction/urban_night_rittenhouse/image/000151.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000151.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 246.5, "y1": 197.0, "x2": 312.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000859" }, { "image": "data/prediction/urban_night_rittenhouse/image/000151.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000151.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 246.5, "y1": 197.0, "x2": 312.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000859" }, { "image": "data/prediction/urban_night_rittenhouse/image/000152.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000152.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 215.0, "y1": 196.5, "x2": 296.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000860" }, { "image": "data/prediction/urban_night_rittenhouse/image/000152.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000152.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 215.0, "y1": 196.5, "x2": 296.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000860" }, { "image": "data/prediction/urban_night_rittenhouse/image/000153.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000153.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 166.0, "y1": 197.0, "x2": 274.0, "y2": 276.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000861" }, { "image": "data/prediction/urban_night_rittenhouse/image/000153.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000153.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 166.0, "y1": 197.0, "x2": 274.0, "y2": 276.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000861" }, { "image": "data/prediction/urban_night_rittenhouse/image/000154.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000154.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 26.5, "y1": 212.0, "x2": 233.0, "y2": 363.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000862" }, { "image": "data/prediction/urban_night_rittenhouse/image/000154.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000154.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 26.5, "y1": 212.0, "x2": 233.0, "y2": 363.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000862" }, { "image": "data/prediction/urban_night_rittenhouse/image/000155.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000155.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 183.0, "y1": 205.0, "x2": 275.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000863" }, { "image": "data/prediction/urban_night_rittenhouse/image/000155.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000155.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 183.0, "y1": 205.0, "x2": 275.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000863" }, { "image": "data/prediction/urban_night_rittenhouse/image/000156.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000156.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 60.0, "y1": 216.0, "x2": 239.0, "y2": 348.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000864" }, { "image": "data/prediction/urban_night_rittenhouse/image/000156.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000156.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 60.0, "y1": 216.0, "x2": 239.0, "y2": 348.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000864" }, { "image": "data/prediction/urban_night_rittenhouse/image/000157.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000157.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 254.5, "y1": 191.5, "x2": 305.0, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000865" }, { "image": "data/prediction/urban_night_rittenhouse/image/000157.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000157.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 254.5, "y1": 191.5, "x2": 305.0, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000865" }, { "image": "data/prediction/urban_night_rittenhouse/image/000158.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000158.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 229.0, "y1": 194.5, "x2": 292.5, "y2": 242.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000866" }, { "image": "data/prediction/urban_night_rittenhouse/image/000158.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000158.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 229.0, "y1": 194.5, "x2": 292.5, "y2": 242.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000866" }, { "image": "data/prediction/urban_night_rittenhouse/image/000159.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000159.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 192.5, "y1": 198.5, "x2": 274.5, "y2": 267.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000867" }, { "image": "data/prediction/urban_night_rittenhouse/image/000159.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000159.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 192.5, "y1": 198.5, "x2": 274.5, "y2": 267.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000867" }, { "image": "data/prediction/urban_night_rittenhouse/image/000160.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000160.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 106.5, "y1": 205.0, "x2": 254.0, "y2": 315.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000868" }, { "image": "data/prediction/urban_night_rittenhouse/image/000160.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000160.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 106.5, "y1": 205.0, "x2": 254.0, "y2": 315.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000868" }, { "image": "data/prediction/urban_night_rittenhouse/image/000161.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000161.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 240.0, "y1": 195.0, "x2": 290.0, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000869" }, { "image": "data/prediction/urban_night_rittenhouse/image/000161.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000161.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 240.0, "y1": 195.0, "x2": 290.0, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000869" }, { "image": "data/prediction/urban_night_rittenhouse/image/000162.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000162.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 201.5, "y1": 198.5, "x2": 286.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000870" }, { "image": "data/prediction/urban_night_rittenhouse/image/000162.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000162.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 201.5, "y1": 198.5, "x2": 286.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000870" }, { "image": "data/prediction/urban_night_rittenhouse/image/000163.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000163.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 138.0, "y1": 204.5, "x2": 262.5, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000871" }, { "image": "data/prediction/urban_night_rittenhouse/image/000163.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000163.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 138.0, "y1": 204.5, "x2": 262.5, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000871" }, { "image": "data/prediction/urban_night_rittenhouse/image/000164.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000164.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 220.0, "x2": 218.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000872" }, { "image": "data/prediction/urban_night_rittenhouse/image/000164.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000164.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 220.0, "x2": 218.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000872" }, { "image": "data/prediction/urban_night_rittenhouse/image/000165.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000165.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 221.0, "x2": 196.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000873" }, { "image": "data/prediction/urban_night_rittenhouse/image/000165.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000165.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 221.0, "x2": 196.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000873" }, { "image": "data/prediction/urban_night_rittenhouse/image/000166.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000166.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 558.0, "y1": 207.5, "x2": 639.5, "y2": 287.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000874" }, { "image": "data/prediction/urban_night_rittenhouse/image/000166.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000166.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 558.0, "y1": 207.5, "x2": 639.5, "y2": 287.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000874" }, { "image": "data/prediction/urban_night_rittenhouse/image/000167.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000167.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 425.5, "y1": 208.0, "x2": 619.5, "y2": 280.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000875" }, { "image": "data/prediction/urban_night_rittenhouse/image/000167.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000167.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 425.5, "y1": 208.0, "x2": 619.5, "y2": 280.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000875" }, { "image": "data/prediction/urban_night_rittenhouse/image/000168.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000168.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 293.5, "y1": 205.5, "x2": 476.5, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000876" }, { "image": "data/prediction/urban_night_rittenhouse/image/000168.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000168.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 293.5, "y1": 205.5, "x2": 476.5, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000876" }, { "image": "data/prediction/urban_night_rittenhouse/image/000169.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000169.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 247.0, "y1": 200.5, "x2": 369.5, "y2": 260.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000877" }, { "image": "data/prediction/urban_night_rittenhouse/image/000169.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000169.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 247.0, "y1": 200.5, "x2": 369.5, "y2": 260.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000877" }, { "image": "data/prediction/urban_night_rittenhouse/image/000170.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000170.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 248.5, "y1": 196.0, "x2": 323.0, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000878" }, { "image": "data/prediction/urban_night_rittenhouse/image/000170.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000170.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 248.5, "y1": 196.0, "x2": 323.0, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000878" }, { "image": "data/prediction/urban_night_rittenhouse/image/000171.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000171.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 591.0, "y1": 195.0, "x2": 639.5, "y2": 253.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000879" }, { "image": "data/prediction/urban_night_rittenhouse/image/000171.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000171.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 591.0, "y1": 195.0, "x2": 639.5, "y2": 253.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000879" }, { "image": "data/prediction/urban_night_rittenhouse/image/000172.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000172.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 499.0, "y1": 192.0, "x2": 639.5, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000880" }, { "image": "data/prediction/urban_night_rittenhouse/image/000172.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000172.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 499.0, "y1": 192.0, "x2": 639.5, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000880" }, { "image": "data/prediction/urban_night_rittenhouse/image/000173.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000173.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 402.0, "y1": 194.5, "x2": 554.5, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000881" }, { "image": "data/prediction/urban_night_rittenhouse/image/000173.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000173.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 402.0, "y1": 194.5, "x2": 554.5, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000881" }, { "image": "data/prediction/urban_night_rittenhouse/image/000174.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000174.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 604.5, "y1": 206.5, "x2": 639.5, "y2": 279.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000882" }, { "image": "data/prediction/urban_night_rittenhouse/image/000174.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000174.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 604.5, "y1": 206.5, "x2": 639.5, "y2": 279.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000882" }, { "image": "data/prediction/urban_night_rittenhouse/image/000175.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000175.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 439.0, "y1": 201.0, "x2": 635.5, "y2": 271.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000883" }, { "image": "data/prediction/urban_night_rittenhouse/image/000175.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000175.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 439.0, "y1": 201.0, "x2": 635.5, "y2": 271.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000883" }, { "image": "data/prediction/urban_night_rittenhouse/image/000176.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000176.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 305.0, "y1": 198.5, "x2": 479.5, "y2": 265.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000884" }, { "image": "data/prediction/urban_night_rittenhouse/image/000176.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000176.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 305.0, "y1": 198.5, "x2": 479.5, "y2": 265.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000884" }, { "image": "data/prediction/urban_night_rittenhouse/image/000177.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000177.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 266.5, "y1": 194.0, "x2": 386.0, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000885" }, { "image": "data/prediction/urban_night_rittenhouse/image/000177.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000177.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 266.5, "y1": 194.0, "x2": 386.0, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000885" }, { "image": "data/prediction/urban_night_rittenhouse/image/000178.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000178.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 271.5, "y1": 192.0, "x2": 339.0, "y2": 241.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000886" }, { "image": "data/prediction/urban_night_rittenhouse/image/000178.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000178.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 271.5, "y1": 192.0, "x2": 339.0, "y2": 241.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000886" }, { "image": "data/prediction/urban_night_rittenhouse/image/000179.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000179.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 595.0, "y1": 196.0, "x2": 639.5, "y2": 255.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000887" }, { "image": "data/prediction/urban_night_rittenhouse/image/000179.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000179.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 595.0, "y1": 196.0, "x2": 639.5, "y2": 255.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000887" }, { "image": "data/prediction/urban_night_rittenhouse/image/000180.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000180.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 484.5, "y1": 195.5, "x2": 639.5, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000888" }, { "image": "data/prediction/urban_night_rittenhouse/image/000180.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000180.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 484.5, "y1": 195.5, "x2": 639.5, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000888" }, { "image": "data/prediction/urban_night_rittenhouse/image/000181.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000181.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 363.0, "y1": 195.5, "x2": 526.0, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000889" }, { "image": "data/prediction/urban_night_rittenhouse/image/000181.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000181.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 363.0, "y1": 195.5, "x2": 526.0, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000889" }, { "image": "data/prediction/urban_night_rittenhouse/image/000182.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000182.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 294.5, "y1": 192.5, "x2": 435.0, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000890" }, { "image": "data/prediction/urban_night_rittenhouse/image/000182.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000182.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 294.5, "y1": 192.5, "x2": 435.0, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000890" }, { "image": "data/prediction/urban_night_rittenhouse/image/000183.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000183.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 279.5, "y1": 191.0, "x2": 371.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000891" }, { "image": "data/prediction/urban_night_rittenhouse/image/000183.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000183.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 279.5, "y1": 191.0, "x2": 371.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000891" }, { "image": "data/prediction/urban_night_rittenhouse/image/000184.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000184.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 287.0, "y1": 190.0, "x2": 339.5, "y2": 233.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000892" }, { "image": "data/prediction/urban_night_rittenhouse/image/000184.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000184.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 287.0, "y1": 190.0, "x2": 339.5, "y2": 233.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000892" }, { "image": "data/prediction/urban_night_rittenhouse/image/000185.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000185.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 319.5, "y1": 193.5, "x2": 434.5, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000893" }, { "image": "data/prediction/urban_night_rittenhouse/image/000185.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000185.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 319.5, "y1": 193.5, "x2": 434.5, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000893" }, { "image": "data/prediction/urban_night_rittenhouse/image/000186.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000186.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 594.0, "y1": 202.0, "x2": 639.5, "y2": 261.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000894" }, { "image": "data/prediction/urban_night_rittenhouse/image/000186.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000186.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 594.0, "y1": 202.0, "x2": 639.5, "y2": 261.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000894" }, { "image": "data/prediction/urban_night_rittenhouse/image/000187.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000187.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 485.0, "y1": 201.0, "x2": 637.0, "y2": 259.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000895" }, { "image": "data/prediction/urban_night_rittenhouse/image/000187.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000187.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 485.0, "y1": 201.0, "x2": 637.0, "y2": 259.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000895" }, { "image": "data/prediction/urban_night_rittenhouse/image/000188.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000188.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 369.0, "y1": 201.5, "x2": 522.0, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000896" }, { "image": "data/prediction/urban_night_rittenhouse/image/000188.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000188.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 369.0, "y1": 201.5, "x2": 522.0, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000896" }, { "image": "data/prediction/urban_night_rittenhouse/image/000189.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000189.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 307.0, "y1": 196.5, "x2": 420.0, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000897" }, { "image": "data/prediction/urban_night_rittenhouse/image/000189.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000189.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 307.0, "y1": 196.5, "x2": 420.0, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000897" }, { "image": "data/prediction/urban_night_rittenhouse/image/000190.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000190.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 285.5, "y1": 196.0, "x2": 358.5, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000898" }, { "image": "data/prediction/urban_night_rittenhouse/image/000190.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000190.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 285.5, "y1": 196.0, "x2": 358.5, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000898" }, { "image": "data/prediction/urban_night_rittenhouse/image/000191.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000191.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 514.5, "y1": 233.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000899" }, { "image": "data/prediction/urban_night_rittenhouse/image/000191.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000191.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 514.5, "y1": 233.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000899" }, { "image": "data/prediction/urban_night_rittenhouse/image/000192.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000192.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 470.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000900" }, { "image": "data/prediction/urban_night_rittenhouse/image/000192.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000192.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 470.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000900" }, { "image": "data/prediction/urban_night_rittenhouse/image/000193.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000193.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 465.5, "y1": 231.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000901" }, { "image": "data/prediction/urban_night_rittenhouse/image/000193.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000193.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 231.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000901" }, { "image": "data/prediction/urban_night_rittenhouse/image/000194.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000194.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000902" }, { "image": "data/prediction/urban_night_rittenhouse/image/000194.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000194.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000902" }, { "image": "data/prediction/urban_night_rittenhouse/image/000195.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000195.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000903" }, { "image": "data/prediction/urban_night_rittenhouse/image/000195.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000195.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000903" }, { "image": "data/prediction/urban_night_rittenhouse/image/000196.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000196.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.5, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000904" }, { "image": "data/prediction/urban_night_rittenhouse/image/000196.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000196.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000904" }, { "image": "data/prediction/urban_night_rittenhouse/image/000197.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000197.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000905" }, { "image": "data/prediction/urban_night_rittenhouse/image/000197.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000197.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000905" }, { "image": "data/prediction/urban_night_rittenhouse/image/000198.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000198.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.5, "y1": 229.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000906" }, { "image": "data/prediction/urban_night_rittenhouse/image/000198.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000198.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.5, "y1": 229.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000906" }, { "image": "data/prediction/urban_night_rittenhouse/image/000199.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000199.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000907" }, { "image": "data/prediction/urban_night_rittenhouse/image/000199.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000199.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000907" }, { "image": "data/prediction/urban_night_rittenhouse/image/000200.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000200.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000908" }, { "image": "data/prediction/urban_night_rittenhouse/image/000200.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000200.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000908" }, { "image": "data/prediction/urban_night_rittenhouse/image/000201.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000201.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000909" }, { "image": "data/prediction/urban_night_rittenhouse/image/000201.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000201.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000909" }, { "image": "data/prediction/urban_night_rittenhouse/image/000202.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000202.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000910" }, { "image": "data/prediction/urban_night_rittenhouse/image/000202.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000202.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000910" }, { "image": "data/prediction/urban_night_rittenhouse/image/000203.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000203.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 461.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000911" }, { "image": "data/prediction/urban_night_rittenhouse/image/000203.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000203.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 461.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000911" }, { "image": "data/prediction/urban_night_rittenhouse/image/000204.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000204.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.0, "y1": 229.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000912" }, { "image": "data/prediction/urban_night_rittenhouse/image/000204.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000204.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.0, "y1": 229.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000912" }, { "image": "data/prediction/urban_night_rittenhouse/image/000205.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000205.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 528.5, "y1": 201.5, "x2": 639.5, "y2": 260.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000913" }, { "image": "data/prediction/urban_night_rittenhouse/image/000205.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000205.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 528.5, "y1": 201.5, "x2": 639.5, "y2": 260.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000913" }, { "image": "data/prediction/urban_night_rittenhouse/image/000206.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000206.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 382.0, "y1": 202.5, "x2": 540.5, "y2": 260.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000914" }, { "image": "data/prediction/urban_night_rittenhouse/image/000206.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000206.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 382.0, "y1": 202.5, "x2": 540.5, "y2": 260.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000914" }, { "image": "data/prediction/urban_night_rittenhouse/image/000207.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000207.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 303.5, "y1": 198.5, "x2": 419.5, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000915" }, { "image": "data/prediction/urban_night_rittenhouse/image/000207.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000207.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 303.5, "y1": 198.5, "x2": 419.5, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000915" }, { "image": "data/prediction/urban_night_rittenhouse/image/000208.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000208.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 290.5, "y1": 194.5, "x2": 352.0, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000916" }, { "image": "data/prediction/urban_night_rittenhouse/image/000208.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000208.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 290.5, "y1": 194.5, "x2": 352.0, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000916" }, { "image": "data/prediction/urban_night_rittenhouse/image/000209.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000209.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 227.0, "y1": 189.5, "x2": 279.0, "y2": 209.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000917" }, { "image": "data/prediction/urban_night_rittenhouse/image/000209.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000209.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 227.0, "y1": 189.5, "x2": 279.0, "y2": 209.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000917" }, { "image": "data/prediction/urban_night_rittenhouse/image/000210.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000210.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 281.0, "y1": 187.5, "x2": 332.0, "y2": 215.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000918" }, { "image": "data/prediction/urban_night_rittenhouse/image/000210.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000210.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 281.0, "y1": 187.5, "x2": 332.0, "y2": 215.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000918" }, { "image": "data/prediction/urban_night_rittenhouse/image/000211.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000211.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 615.5, "y1": 195.0, "x2": 639.5, "y2": 254.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000919" }, { "image": "data/prediction/urban_night_rittenhouse/image/000211.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000211.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 615.5, "y1": 195.0, "x2": 639.5, "y2": 254.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000919" }, { "image": "data/prediction/urban_night_rittenhouse/image/000212.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000212.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 487.0, "y1": 192.5, "x2": 637.5, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000920" }, { "image": "data/prediction/urban_night_rittenhouse/image/000212.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000212.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 487.0, "y1": 192.5, "x2": 637.5, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000920" }, { "image": "data/prediction/urban_night_rittenhouse/image/000213.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000213.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 341.5, "y1": 191.5, "x2": 456.0, "y2": 243.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000921" }, { "image": "data/prediction/urban_night_rittenhouse/image/000213.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000213.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 341.5, "y1": 191.5, "x2": 456.0, "y2": 243.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000921" }, { "image": "data/prediction/urban_night_rittenhouse/image/000214.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000214.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 333.5, "y1": 191.0, "x2": 396.5, "y2": 233.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000922" }, { "image": "data/prediction/urban_night_rittenhouse/image/000214.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000214.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 333.5, "y1": 191.0, "x2": 396.5, "y2": 233.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000922" }, { "image": "data/prediction/urban_night_rittenhouse/image/000215.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000215.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 463.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000923" }, { "image": "data/prediction/urban_night_rittenhouse/image/000215.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000215.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 463.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000923" }, { "image": "data/prediction/urban_night_rittenhouse/image/000216.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000216.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.0, "y1": 229.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000924" }, { "image": "data/prediction/urban_night_rittenhouse/image/000216.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000216.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.0, "y1": 229.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000924" }, { "image": "data/prediction/urban_night_rittenhouse/image/000217.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000217.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000925" }, { "image": "data/prediction/urban_night_rittenhouse/image/000217.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000217.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.0, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000925" }, { "image": "data/prediction/urban_night_rittenhouse/image/000218.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000218.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.5, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000926" }, { "image": "data/prediction/urban_night_rittenhouse/image/000218.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000218.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 230.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000926" }, { "image": "data/prediction/urban_night_rittenhouse/image/000219.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000219.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 464.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000927" }, { "image": "data/prediction/urban_night_rittenhouse/image/000219.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000219.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 464.5, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000927" }, { "image": "data/prediction/urban_night_rittenhouse/image/000220.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000220.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 458.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000928" }, { "image": "data/prediction/urban_night_rittenhouse/image/000220.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000220.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 458.0, "y1": 231.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000928" }, { "image": "data/prediction/urban_night_rittenhouse/image/000221.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000221.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 467.5, "y1": 229.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000929" }, { "image": "data/prediction/urban_night_rittenhouse/image/000221.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000221.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 467.5, "y1": 229.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000929" }, { "image": "data/prediction/urban_night_rittenhouse/image/000222.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000222.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 461.5, "y1": 223.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000930" }, { "image": "data/prediction/urban_night_rittenhouse/image/000222.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000222.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 461.5, "y1": 223.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000930" }, { "image": "data/prediction/urban_night_rittenhouse/image/000223.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000223.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 439.5, "y1": 216.5, "x2": 639.5, "y2": 387.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000931" }, { "image": "data/prediction/urban_night_rittenhouse/image/000223.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000223.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 439.5, "y1": 216.5, "x2": 639.5, "y2": 387.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000931" }, { "image": "data/prediction/urban_night_rittenhouse/image/000224.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000224.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 428.5, "y1": 215.0, "x2": 596.0, "y2": 338.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000932" }, { "image": "data/prediction/urban_night_rittenhouse/image/000224.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000224.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.5, "y1": 215.0, "x2": 596.0, "y2": 338.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000932" }, { "image": "data/prediction/urban_night_rittenhouse/image/000225.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000225.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 424.5, "y1": 211.5, "x2": 558.0, "y2": 309.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000933" }, { "image": "data/prediction/urban_night_rittenhouse/image/000225.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000225.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.5, "y1": 211.5, "x2": 558.0, "y2": 309.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000933" }, { "image": "data/prediction/urban_night_rittenhouse/image/000226.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000226.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 414.5, "y1": 205.0, "x2": 520.5, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000934" }, { "image": "data/prediction/urban_night_rittenhouse/image/000226.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000226.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.5, "y1": 205.0, "x2": 520.5, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000934" }, { "image": "data/prediction/urban_night_rittenhouse/image/000227.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000227.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 388.5, "y1": 199.0, "x2": 482.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000935" }, { "image": "data/prediction/urban_night_rittenhouse/image/000227.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000227.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 388.5, "y1": 199.0, "x2": 482.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000935" }, { "image": "data/prediction/urban_night_rittenhouse/image/000228.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000228.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 347.5, "y1": 190.5, "x2": 426.0, "y2": 242.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000936" }, { "image": "data/prediction/urban_night_rittenhouse/image/000228.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000228.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 347.5, "y1": 190.5, "x2": 426.0, "y2": 242.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000936" }, { "image": "data/prediction/urban_night_rittenhouse/image/000229.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000229.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 346.5, "y1": 196.5, "x2": 397.5, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000937" }, { "image": "data/prediction/urban_night_rittenhouse/image/000229.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000229.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 346.5, "y1": 196.5, "x2": 397.5, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000937" }, { "image": "data/prediction/urban_night_rittenhouse/image/000230.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000230.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 179.0, "y1": 185.0, "x2": 237.5, "y2": 223.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000938" }, { "image": "data/prediction/urban_night_rittenhouse/image/000230.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000230.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 179.0, "y1": 185.0, "x2": 237.5, "y2": 223.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000938" }, { "image": "data/prediction/urban_night_rittenhouse/image/000231.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000231.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 336.0, "y1": 203.0, "x2": 405.0, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000939" }, { "image": "data/prediction/urban_night_rittenhouse/image/000231.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000231.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 336.0, "y1": 203.0, "x2": 405.0, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000939" }, { "image": "data/prediction/urban_night_rittenhouse/image/000232.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000232.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 202.0, "y1": 182.5, "x2": 253.5, "y2": 216.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000940" }, { "image": "data/prediction/urban_night_rittenhouse/image/000232.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000232.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 202.0, "y1": 182.5, "x2": 253.5, "y2": 216.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000940" }, { "image": "data/prediction/urban_night_rittenhouse/image/000233.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000233.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 341.5, "y1": 198.0, "x2": 400.5, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000941" }, { "image": "data/prediction/urban_night_rittenhouse/image/000233.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000233.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 341.5, "y1": 198.0, "x2": 400.5, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000941" }, { "image": "data/prediction/urban_night_rittenhouse/image/000234.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000234.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 410.0, "y1": 202.5, "x2": 496.0, "y2": 258.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000942" }, { "image": "data/prediction/urban_night_rittenhouse/image/000234.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000234.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 410.0, "y1": 202.5, "x2": 496.0, "y2": 258.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000942" }, { "image": "data/prediction/urban_night_rittenhouse/image/000235.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000235.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 461.0, "y1": 208.5, "x2": 592.5, "y2": 293.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000943" }, { "image": "data/prediction/urban_night_rittenhouse/image/000235.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000235.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 461.0, "y1": 208.5, "x2": 592.5, "y2": 293.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000943" }, { "image": "data/prediction/urban_night_rittenhouse/image/000236.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000236.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 442.5, "y1": 203.0, "x2": 547.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000944" }, { "image": "data/prediction/urban_night_rittenhouse/image/000236.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000236.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 442.5, "y1": 203.0, "x2": 547.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000944" }, { "image": "data/prediction/urban_night_rittenhouse/image/000237.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000237.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 364.0, "y1": 194.0, "x2": 429.0, "y2": 230.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000945" }, { "image": "data/prediction/urban_night_rittenhouse/image/000237.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000237.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 364.0, "y1": 194.0, "x2": 429.0, "y2": 230.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000945" }, { "image": "data/prediction/urban_night_rittenhouse/image/000238.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000238.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 408.0, "y1": 200.5, "x2": 498.5, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000946" }, { "image": "data/prediction/urban_night_rittenhouse/image/000238.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000238.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.0, "y1": 200.5, "x2": 498.5, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000946" }, { "image": "data/prediction/urban_night_rittenhouse/image/000239.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000239.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 477.5, "y1": 201.5, "x2": 599.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000947" }, { "image": "data/prediction/urban_night_rittenhouse/image/000239.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000239.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 477.5, "y1": 201.5, "x2": 599.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000947" }, { "image": "data/prediction/urban_night_rittenhouse/image/000240.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000240.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 340.5, "y1": 194.5, "x2": 394.5, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000948" }, { "image": "data/prediction/urban_night_rittenhouse/image/000240.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000240.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 340.5, "y1": 194.5, "x2": 394.5, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000948" }, { "image": "data/prediction/urban_night_rittenhouse/image/000241.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000241.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 342.0, "y1": 197.5, "x2": 407.0, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000949" }, { "image": "data/prediction/urban_night_rittenhouse/image/000241.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000241.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 342.0, "y1": 197.5, "x2": 407.0, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000949" }, { "image": "data/prediction/urban_night_rittenhouse/image/000242.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000242.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 339.0, "y1": 203.0, "x2": 420.0, "y2": 275.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000950" }, { "image": "data/prediction/urban_night_rittenhouse/image/000242.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000242.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 339.0, "y1": 203.0, "x2": 420.0, "y2": 275.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000950" }, { "image": "data/prediction/urban_night_rittenhouse/image/000243.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000243.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 344.0, "y1": 202.0, "x2": 447.5, "y2": 295.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000951" }, { "image": "data/prediction/urban_night_rittenhouse/image/000243.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000243.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 344.0, "y1": 202.0, "x2": 447.5, "y2": 295.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000951" }, { "image": "data/prediction/urban_night_rittenhouse/image/000244.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000244.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 407.5, "y1": 193.0, "x2": 461.0, "y2": 230.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000952" }, { "image": "data/prediction/urban_night_rittenhouse/image/000244.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000244.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 407.5, "y1": 193.0, "x2": 461.0, "y2": 230.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000952" }, { "image": "data/prediction/urban_night_rittenhouse/image/000245.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000245.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 421.0, "y1": 195.5, "x2": 483.0, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000953" }, { "image": "data/prediction/urban_night_rittenhouse/image/000245.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000245.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.0, "y1": 195.5, "x2": 483.0, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000953" }, { "image": "data/prediction/urban_night_rittenhouse/image/000246.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000246.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 431.0, "y1": 197.0, "x2": 509.5, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000954" }, { "image": "data/prediction/urban_night_rittenhouse/image/000246.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000246.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.0, "y1": 197.0, "x2": 509.5, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000954" }, { "image": "data/prediction/urban_night_rittenhouse/image/000247.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000247.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 438.0, "y1": 197.0, "x2": 538.0, "y2": 262.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000955" }, { "image": "data/prediction/urban_night_rittenhouse/image/000247.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000247.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 438.0, "y1": 197.0, "x2": 538.0, "y2": 262.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000955" }, { "image": "data/prediction/urban_night_rittenhouse/image/000248.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000248.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 411.5, "y1": 197.0, "x2": 478.5, "y2": 242.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000956" }, { "image": "data/prediction/urban_night_rittenhouse/image/000248.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000248.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.5, "y1": 197.0, "x2": 478.5, "y2": 242.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000956" }, { "image": "data/prediction/urban_night_rittenhouse/image/000249.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000249.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 422.5, "y1": 201.5, "x2": 509.0, "y2": 253.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000957" }, { "image": "data/prediction/urban_night_rittenhouse/image/000249.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000249.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 422.5, "y1": 201.5, "x2": 509.0, "y2": 253.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000957" }, { "image": "data/prediction/urban_night_rittenhouse/image/000250.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000250.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 441.5, "y1": 200.5, "x2": 531.0, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000958" }, { "image": "data/prediction/urban_night_rittenhouse/image/000250.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000250.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.5, "y1": 200.5, "x2": 531.0, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000958" }, { "image": "data/prediction/urban_night_rittenhouse/image/000251.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000251.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 440.5, "y1": 201.0, "x2": 534.5, "y2": 259.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000959" }, { "image": "data/prediction/urban_night_rittenhouse/image/000251.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000251.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 440.5, "y1": 201.0, "x2": 534.5, "y2": 259.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000959" }, { "image": "data/prediction/urban_night_rittenhouse/image/000252.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000252.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 413.5, "y1": 202.5, "x2": 514.0, "y2": 264.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000960" }, { "image": "data/prediction/urban_night_rittenhouse/image/000252.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000252.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 413.5, "y1": 202.5, "x2": 514.0, "y2": 264.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000960" }, { "image": "data/prediction/urban_night_rittenhouse/image/000253.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000253.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.5, "y1": 201.0, "x2": 509.0, "y2": 265.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000961" }, { "image": "data/prediction/urban_night_rittenhouse/image/000253.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000253.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.5, "y1": 201.0, "x2": 509.0, "y2": 265.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000961" }, { "image": "data/prediction/urban_night_rittenhouse/image/000254.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000254.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 412.5, "y1": 201.5, "x2": 517.5, "y2": 269.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000962" }, { "image": "data/prediction/urban_night_rittenhouse/image/000254.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000254.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.5, "y1": 201.5, "x2": 517.5, "y2": 269.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000962" }, { "image": "data/prediction/urban_night_rittenhouse/image/000255.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000255.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 414.5, "y1": 202.5, "x2": 516.0, "y2": 270.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000963" }, { "image": "data/prediction/urban_night_rittenhouse/image/000255.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000255.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.5, "y1": 202.5, "x2": 516.0, "y2": 270.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000963" }, { "image": "data/prediction/urban_night_rittenhouse/image/000256.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000256.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 416.0, "y1": 203.0, "x2": 518.5, "y2": 271.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000964" }, { "image": "data/prediction/urban_night_rittenhouse/image/000256.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000256.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.0, "y1": 203.0, "x2": 518.5, "y2": 271.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000964" }, { "image": "data/prediction/urban_night_rittenhouse/image/000257.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000257.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.5, "y1": 203.0, "x2": 517.5, "y2": 272.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000965" }, { "image": "data/prediction/urban_night_rittenhouse/image/000257.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000257.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.5, "y1": 203.0, "x2": 517.5, "y2": 272.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000965" }, { "image": "data/prediction/urban_night_rittenhouse/image/000258.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000258.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 416.0, "y1": 204.0, "x2": 518.5, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000966" }, { "image": "data/prediction/urban_night_rittenhouse/image/000258.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000258.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.0, "y1": 204.0, "x2": 518.5, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000966" }, { "image": "data/prediction/urban_night_rittenhouse/image/000259.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000259.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 343.5, "y1": 191.0, "x2": 394.5, "y2": 225.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000967" }, { "image": "data/prediction/urban_night_rittenhouse/image/000259.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000259.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 343.5, "y1": 191.0, "x2": 394.5, "y2": 225.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000967" }, { "image": "data/prediction/urban_night_rittenhouse/image/000260.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000260.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 303.0, "y1": 191.0, "x2": 356.0, "y2": 226.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000968" }, { "image": "data/prediction/urban_night_rittenhouse/image/000260.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000260.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 303.0, "y1": 191.0, "x2": 356.0, "y2": 226.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000968" }, { "image": "data/prediction/urban_night_rittenhouse/image/000261.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000261.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 456.5, "y1": 198.5, "x2": 568.0, "y2": 272.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000969" }, { "image": "data/prediction/urban_night_rittenhouse/image/000261.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000261.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 456.5, "y1": 198.5, "x2": 568.0, "y2": 272.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000969" }, { "image": "data/prediction/urban_night_rittenhouse/image/000262.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000262.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 475.5, "y1": 199.0, "x2": 599.0, "y2": 280.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000970" }, { "image": "data/prediction/urban_night_rittenhouse/image/000262.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000262.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 475.5, "y1": 199.0, "x2": 599.0, "y2": 280.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000970" }, { "image": "data/prediction/urban_night_rittenhouse/image/000263.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000263.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 484.5, "y1": 201.5, "x2": 618.5, "y2": 289.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000971" }, { "image": "data/prediction/urban_night_rittenhouse/image/000263.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000263.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 484.5, "y1": 201.5, "x2": 618.5, "y2": 289.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000971" }, { "image": "data/prediction/urban_night_rittenhouse/image/000264.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000264.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.0, "y1": 201.5, "x2": 601.0, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000972" }, { "image": "data/prediction/urban_night_rittenhouse/image/000264.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000264.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.0, "y1": 201.5, "x2": 601.0, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000972" }, { "image": "data/prediction/urban_night_rittenhouse/image/000265.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000265.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 448.0, "y1": 202.0, "x2": 611.0, "y2": 308.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000973" }, { "image": "data/prediction/urban_night_rittenhouse/image/000265.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000265.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 448.0, "y1": 202.0, "x2": 611.0, "y2": 308.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000973" }, { "image": "data/prediction/urban_night_rittenhouse/image/000266.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000266.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.0, "y1": 203.5, "x2": 628.5, "y2": 316.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000974" }, { "image": "data/prediction/urban_night_rittenhouse/image/000266.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000266.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.0, "y1": 203.5, "x2": 628.5, "y2": 316.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000974" }, { "image": "data/prediction/urban_night_rittenhouse/image/000267.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000267.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 454.5, "y1": 202.5, "x2": 631.0, "y2": 317.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000975" }, { "image": "data/prediction/urban_night_rittenhouse/image/000267.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000267.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 454.5, "y1": 202.5, "x2": 631.0, "y2": 317.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000975" }, { "image": "data/prediction/urban_night_rittenhouse/image/000268.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000268.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 454.5, "y1": 203.0, "x2": 629.5, "y2": 318.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000976" }, { "image": "data/prediction/urban_night_rittenhouse/image/000268.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000268.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 454.5, "y1": 203.0, "x2": 629.5, "y2": 318.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000976" }, { "image": "data/prediction/urban_night_rittenhouse/image/000269.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000269.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 455.0, "y1": 203.0, "x2": 630.5, "y2": 317.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000977" }, { "image": "data/prediction/urban_night_rittenhouse/image/000269.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000269.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 455.0, "y1": 203.0, "x2": 630.5, "y2": 317.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000977" }, { "image": "data/prediction/urban_night_rittenhouse/image/000270.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000270.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 455.0, "y1": 205.0, "x2": 632.0, "y2": 319.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000978" }, { "image": "data/prediction/urban_night_rittenhouse/image/000270.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000270.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 455.0, "y1": 205.0, "x2": 632.0, "y2": 319.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000978" }, { "image": "data/prediction/urban_night_rittenhouse/image/000271.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000271.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 346.0, "y1": 206.0, "x2": 477.5, "y2": 321.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000979" }, { "image": "data/prediction/urban_night_rittenhouse/image/000271.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000271.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 346.0, "y1": 206.0, "x2": 477.5, "y2": 321.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000979" }, { "image": "data/prediction/urban_night_rittenhouse/image/000272.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000272.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.0, "y1": 209.5, "x2": 508.5, "y2": 341.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000980" }, { "image": "data/prediction/urban_night_rittenhouse/image/000272.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000272.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.0, "y1": 209.5, "x2": 508.5, "y2": 341.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000980" }, { "image": "data/prediction/urban_night_rittenhouse/image/000273.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000273.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 356.5, "y1": 212.5, "x2": 532.5, "y2": 365.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000981" }, { "image": "data/prediction/urban_night_rittenhouse/image/000273.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000273.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 212.5, "x2": 532.5, "y2": 365.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000981" }, { "image": "data/prediction/urban_night_rittenhouse/image/000274.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000274.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 314.0, "y1": 215.5, "x2": 521.5, "y2": 396.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000982" }, { "image": "data/prediction/urban_night_rittenhouse/image/000274.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000274.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 314.0, "y1": 215.5, "x2": 521.5, "y2": 396.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000982" }, { "image": "data/prediction/urban_night_rittenhouse/image/000275.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000275.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 301.0, "y1": 218.5, "x2": 543.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000983" }, { "image": "data/prediction/urban_night_rittenhouse/image/000275.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000275.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 301.0, "y1": 218.5, "x2": 543.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000983" }, { "image": "data/prediction/urban_night_rittenhouse/image/000276.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000276.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 299.5, "y1": 221.5, "x2": 569.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000984" }, { "image": "data/prediction/urban_night_rittenhouse/image/000276.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000276.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.5, "y1": 221.5, "x2": 569.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000984" }, { "image": "data/prediction/urban_night_rittenhouse/image/000277.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000277.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 299.5, "y1": 221.5, "x2": 572.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000985" }, { "image": "data/prediction/urban_night_rittenhouse/image/000277.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000277.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.5, "y1": 221.5, "x2": 572.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000985" }, { "image": "data/prediction/urban_night_rittenhouse/image/000278.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000278.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 298.0, "y1": 222.0, "x2": 570.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000986" }, { "image": "data/prediction/urban_night_rittenhouse/image/000278.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000278.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 298.0, "y1": 222.0, "x2": 570.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000986" }, { "image": "data/prediction/urban_night_rittenhouse/image/000279.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000279.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 297.0, "y1": 220.5, "x2": 571.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000987" }, { "image": "data/prediction/urban_night_rittenhouse/image/000279.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000279.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 297.0, "y1": 220.5, "x2": 571.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000987" }, { "image": "data/prediction/urban_night_rittenhouse/image/000280.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000280.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 299.5, "y1": 220.5, "x2": 551.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000988" }, { "image": "data/prediction/urban_night_rittenhouse/image/000280.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000280.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.5, "y1": 220.5, "x2": 551.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000988" }, { "image": "data/prediction/urban_night_rittenhouse/image/000281.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000281.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 354.5, "y1": 195.5, "x2": 456.5, "y2": 274.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000989" }, { "image": "data/prediction/urban_night_rittenhouse/image/000281.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000281.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 354.5, "y1": 195.5, "x2": 456.5, "y2": 274.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000989" }, { "image": "data/prediction/urban_night_rittenhouse/image/000282.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000282.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 315.5, "y1": 198.0, "x2": 415.0, "y2": 286.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000990" }, { "image": "data/prediction/urban_night_rittenhouse/image/000282.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000282.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 315.5, "y1": 198.0, "x2": 415.0, "y2": 286.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000990" }, { "image": "data/prediction/urban_night_rittenhouse/image/000283.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000283.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 312.0, "y1": 200.0, "x2": 417.0, "y2": 294.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000991" }, { "image": "data/prediction/urban_night_rittenhouse/image/000283.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000283.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 312.0, "y1": 200.0, "x2": 417.0, "y2": 294.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000991" }, { "image": "data/prediction/urban_night_rittenhouse/image/000284.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000284.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 309.0, "y1": 200.5, "x2": 426.0, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000992" }, { "image": "data/prediction/urban_night_rittenhouse/image/000284.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000284.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.0, "y1": 200.5, "x2": 426.0, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000992" }, { "image": "data/prediction/urban_night_rittenhouse/image/000285.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000285.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 310.0, "y1": 203.5, "x2": 425.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000993" }, { "image": "data/prediction/urban_night_rittenhouse/image/000285.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000285.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 310.0, "y1": 203.5, "x2": 425.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000993" }, { "image": "data/prediction/urban_night_rittenhouse/image/000286.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000286.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 309.5, "y1": 204.0, "x2": 426.0, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000994" }, { "image": "data/prediction/urban_night_rittenhouse/image/000286.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000286.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.5, "y1": 204.0, "x2": 426.0, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000994" }, { "image": "data/prediction/urban_night_rittenhouse/image/000287.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000287.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 310.0, "y1": 202.5, "x2": 425.0, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000995" }, { "image": "data/prediction/urban_night_rittenhouse/image/000287.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000287.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 310.0, "y1": 202.5, "x2": 425.0, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000995" }, { "image": "data/prediction/urban_night_rittenhouse/image/000288.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000288.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 312.5, "y1": 192.0, "x2": 400.5, "y2": 241.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000996" }, { "image": "data/prediction/urban_night_rittenhouse/image/000288.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000288.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 312.5, "y1": 192.0, "x2": 400.5, "y2": 241.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000996" }, { "image": "data/prediction/urban_night_rittenhouse/image/000289.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000289.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 311.0, "y1": 201.0, "x2": 402.5, "y2": 282.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000997" }, { "image": "data/prediction/urban_night_rittenhouse/image/000289.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000289.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 311.0, "y1": 201.0, "x2": 402.5, "y2": 282.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000997" }, { "image": "data/prediction/urban_night_rittenhouse/image/000290.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000290.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 295.5, "y1": 202.0, "x2": 388.0, "y2": 269.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000998" }, { "image": "data/prediction/urban_night_rittenhouse/image/000290.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000290.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 295.5, "y1": 202.0, "x2": 388.0, "y2": 269.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000998" }, { "image": "data/prediction/urban_night_rittenhouse/image/000291.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000291.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.0, "y1": 202.5, "x2": 523.5, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:000999" }, { "image": "data/prediction/urban_night_rittenhouse/image/000291.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000291.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.0, "y1": 202.5, "x2": 523.5, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:000999" }, { "image": "data/prediction/urban_night_rittenhouse/image/000292.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000292.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.5, "y1": 210.0, "x2": 536.0, "y2": 286.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001000" }, { "image": "data/prediction/urban_night_rittenhouse/image/000292.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000292.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.5, "y1": 210.0, "x2": 536.0, "y2": 286.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001000" }, { "image": "data/prediction/urban_night_rittenhouse/image/000293.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000293.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.0, "y1": 213.5, "x2": 563.0, "y2": 299.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001001" }, { "image": "data/prediction/urban_night_rittenhouse/image/000293.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000293.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.0, "y1": 213.5, "x2": 563.0, "y2": 299.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001001" }, { "image": "data/prediction/urban_night_rittenhouse/image/000294.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000294.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.5, "y1": 218.0, "x2": 612.0, "y2": 321.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001002" }, { "image": "data/prediction/urban_night_rittenhouse/image/000294.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000294.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.5, "y1": 218.0, "x2": 612.0, "y2": 321.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001002" }, { "image": "data/prediction/urban_night_rittenhouse/image/000295.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000295.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 504.5, "y1": 221.5, "x2": 639.5, "y2": 367.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001003" }, { "image": "data/prediction/urban_night_rittenhouse/image/000295.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000295.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 504.5, "y1": 221.5, "x2": 639.5, "y2": 367.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001003" }, { "image": "data/prediction/urban_night_rittenhouse/image/000296.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000296.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 458.0, "y1": 205.5, "x2": 638.0, "y2": 323.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001004" }, { "image": "data/prediction/urban_night_rittenhouse/image/000296.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000296.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 458.0, "y1": 205.5, "x2": 638.0, "y2": 323.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001004" }, { "image": "data/prediction/urban_night_rittenhouse/image/000297.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000297.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 469.0, "y1": 210.5, "x2": 639.5, "y2": 339.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001005" }, { "image": "data/prediction/urban_night_rittenhouse/image/000297.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000297.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 469.0, "y1": 210.5, "x2": 639.5, "y2": 339.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001005" }, { "image": "data/prediction/urban_night_rittenhouse/image/000298.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000298.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 491.5, "y1": 216.5, "x2": 639.5, "y2": 368.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001006" }, { "image": "data/prediction/urban_night_rittenhouse/image/000298.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000298.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 491.5, "y1": 216.5, "x2": 639.5, "y2": 368.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001006" }, { "image": "data/prediction/urban_night_rittenhouse/image/000299.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000299.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 529.0, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001007" }, { "image": "data/prediction/urban_night_rittenhouse/image/000299.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000299.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 529.0, "y1": 230.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001007" }, { "image": "data/prediction/urban_night_rittenhouse/image/000300.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000300.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 303.0, "y1": 220.0, "x2": 529.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001008" }, { "image": "data/prediction/urban_night_rittenhouse/image/000300.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000300.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 303.0, "y1": 220.0, "x2": 529.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001008" }, { "image": "data/prediction/urban_night_rittenhouse/image/000301.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000301.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 298.5, "y1": 220.5, "x2": 494.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001009" }, { "image": "data/prediction/urban_night_rittenhouse/image/000301.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000301.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 298.5, "y1": 220.5, "x2": 494.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001009" }, { "image": "data/prediction/urban_night_rittenhouse/image/000302.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000302.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 309.0, "y1": 218.0, "x2": 460.5, "y2": 359.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001010" }, { "image": "data/prediction/urban_night_rittenhouse/image/000302.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000302.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.0, "y1": 218.0, "x2": 460.5, "y2": 359.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001010" }, { "image": "data/prediction/urban_night_rittenhouse/image/000303.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000303.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 319.0, "y1": 216.0, "x2": 437.5, "y2": 327.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001011" }, { "image": "data/prediction/urban_night_rittenhouse/image/000303.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000303.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 319.0, "y1": 216.0, "x2": 437.5, "y2": 327.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001011" }, { "image": "data/prediction/urban_night_rittenhouse/image/000304.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000304.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 324.5, "y1": 212.0, "x2": 421.0, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001012" }, { "image": "data/prediction/urban_night_rittenhouse/image/000304.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000304.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.5, "y1": 212.0, "x2": 421.0, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001012" }, { "image": "data/prediction/urban_night_rittenhouse/image/000305.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000305.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 330.5, "y1": 204.5, "x2": 414.5, "y2": 281.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001013" }, { "image": "data/prediction/urban_night_rittenhouse/image/000305.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000305.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 330.5, "y1": 204.5, "x2": 414.5, "y2": 281.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001013" }, { "image": "data/prediction/urban_night_rittenhouse/image/000306.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000306.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 332.0, "y1": 205.5, "x2": 414.0, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001014" }, { "image": "data/prediction/urban_night_rittenhouse/image/000306.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000306.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 332.0, "y1": 205.5, "x2": 414.0, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001014" }, { "image": "data/prediction/urban_night_rittenhouse/image/000307.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000307.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 330.0, "y1": 205.0, "x2": 413.0, "y2": 280.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001015" }, { "image": "data/prediction/urban_night_rittenhouse/image/000307.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000307.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 330.0, "y1": 205.0, "x2": 413.0, "y2": 280.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001015" }, { "image": "data/prediction/urban_night_rittenhouse/image/000308.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000308.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 323.0, "y1": 203.0, "x2": 404.5, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001016" }, { "image": "data/prediction/urban_night_rittenhouse/image/000308.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000308.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.0, "y1": 203.0, "x2": 404.5, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001016" }, { "image": "data/prediction/urban_night_rittenhouse/image/000309.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000309.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 324.0, "y1": 204.5, "x2": 403.0, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001017" }, { "image": "data/prediction/urban_night_rittenhouse/image/000309.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000309.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.0, "y1": 204.5, "x2": 403.0, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001017" }, { "image": "data/prediction/urban_night_rittenhouse/image/000310.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000310.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 311.5, "y1": 195.0, "x2": 364.5, "y2": 234.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001018" }, { "image": "data/prediction/urban_night_rittenhouse/image/000310.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000310.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 311.5, "y1": 195.0, "x2": 364.5, "y2": 234.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001018" }, { "image": "data/prediction/urban_night_rittenhouse/image/000311.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000311.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 306.5, "y1": 195.5, "x2": 367.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001019" }, { "image": "data/prediction/urban_night_rittenhouse/image/000311.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000311.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 306.5, "y1": 195.5, "x2": 367.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001019" }, { "image": "data/prediction/urban_night_rittenhouse/image/000312.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000312.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 376.0, "y1": 173.0, "x2": 428.5, "y2": 222.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001020" }, { "image": "data/prediction/urban_night_rittenhouse/image/000312.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000312.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 376.0, "y1": 173.0, "x2": 428.5, "y2": 222.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001020" }, { "image": "data/prediction/urban_night_rittenhouse/image/000313.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000313.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 377.5, "y1": 165.5, "x2": 448.0, "y2": 225.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001021" }, { "image": "data/prediction/urban_night_rittenhouse/image/000313.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000313.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 377.5, "y1": 165.5, "x2": 448.0, "y2": 225.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001021" }, { "image": "data/prediction/urban_night_rittenhouse/image/000314.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000314.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 388.0, "y1": 165.5, "x2": 470.5, "y2": 234.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001022" }, { "image": "data/prediction/urban_night_rittenhouse/image/000314.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000314.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 388.0, "y1": 165.5, "x2": 470.5, "y2": 234.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001022" }, { "image": "data/prediction/urban_night_rittenhouse/image/000315.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000315.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 297.0, "y1": 197.0, "x2": 361.5, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001023" }, { "image": "data/prediction/urban_night_rittenhouse/image/000315.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000315.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 297.0, "y1": 197.0, "x2": 361.5, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001023" }, { "image": "data/prediction/urban_night_rittenhouse/image/000316.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000316.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 397.0, "y1": 164.5, "x2": 504.0, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001024" }, { "image": "data/prediction/urban_night_rittenhouse/image/000316.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000316.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 397.0, "y1": 164.5, "x2": 504.0, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001024" }, { "image": "data/prediction/urban_night_rittenhouse/image/000317.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000317.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 413.0, "y1": 160.5, "x2": 564.5, "y2": 281.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001025" }, { "image": "data/prediction/urban_night_rittenhouse/image/000317.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000317.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 413.0, "y1": 160.5, "x2": 564.5, "y2": 281.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001025" }, { "image": "data/prediction/urban_night_rittenhouse/image/000318.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000318.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 441.0, "y1": 149.0, "x2": 639.5, "y2": 334.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001026" }, { "image": "data/prediction/urban_night_rittenhouse/image/000318.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000318.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.0, "y1": 149.0, "x2": 639.5, "y2": 334.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001026" }, { "image": "data/prediction/urban_night_rittenhouse/image/000319.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000319.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 478.5, "y1": 128.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001027" }, { "image": "data/prediction/urban_night_rittenhouse/image/000319.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000319.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 478.5, "y1": 128.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001027" }, { "image": "data/prediction/urban_night_rittenhouse/image/000320.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000320.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 323.0, "y1": 200.0, "x2": 397.5, "y2": 266.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001028" }, { "image": "data/prediction/urban_night_rittenhouse/image/000320.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000320.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 323.0, "y1": 200.0, "x2": 397.5, "y2": 266.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001028" }, { "image": "data/prediction/urban_night_rittenhouse/image/000321.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000321.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 329.5, "y1": 201.5, "x2": 402.5, "y2": 266.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001029" }, { "image": "data/prediction/urban_night_rittenhouse/image/000321.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000321.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 329.5, "y1": 201.5, "x2": 402.5, "y2": 266.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001029" }, { "image": "data/prediction/urban_night_rittenhouse/image/000322.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000322.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 335.5, "y1": 203.5, "x2": 409.0, "y2": 270.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001030" }, { "image": "data/prediction/urban_night_rittenhouse/image/000322.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000322.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 335.5, "y1": 203.5, "x2": 409.0, "y2": 270.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001030" }, { "image": "data/prediction/urban_night_rittenhouse/image/000323.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000323.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 332.5, "y1": 205.0, "x2": 409.5, "y2": 273.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001031" }, { "image": "data/prediction/urban_night_rittenhouse/image/000323.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000323.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 332.5, "y1": 205.0, "x2": 409.5, "y2": 273.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001031" }, { "image": "data/prediction/urban_night_rittenhouse/image/000324.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000324.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 337.5, "y1": 203.0, "x2": 418.0, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001032" }, { "image": "data/prediction/urban_night_rittenhouse/image/000324.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000324.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 337.5, "y1": 203.0, "x2": 418.0, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001032" }, { "image": "data/prediction/urban_night_rittenhouse/image/000325.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000325.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 273.0, "y1": 201.0, "x2": 399.0, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001033" }, { "image": "data/prediction/urban_night_rittenhouse/image/000325.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000325.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 273.0, "y1": 201.0, "x2": 399.0, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001033" }, { "image": "data/prediction/urban_night_rittenhouse/image/000326.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000326.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 152.5, "y1": 198.5, "x2": 307.5, "y2": 283.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001034" }, { "image": "data/prediction/urban_night_rittenhouse/image/000326.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000326.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 152.5, "y1": 198.5, "x2": 307.5, "y2": 283.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001034" }, { "image": "data/prediction/urban_night_rittenhouse/image/000327.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000327.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 197.0, "x2": 131.0, "y2": 286.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001035" }, { "image": "data/prediction/urban_night_rittenhouse/image/000327.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000327.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 197.0, "x2": 131.0, "y2": 286.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001035" }, { "image": "data/prediction/urban_night_rittenhouse/image/000328.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000328.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 203.0, "x2": 55.0, "y2": 293.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001036" }, { "image": "data/prediction/urban_night_rittenhouse/image/000328.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000328.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 203.0, "x2": 55.0, "y2": 293.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001036" }, { "image": "data/prediction/urban_night_rittenhouse/image/000329.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000329.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 201.0, "x2": 111.5, "y2": 284.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001037" }, { "image": "data/prediction/urban_night_rittenhouse/image/000329.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000329.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 201.0, "x2": 111.5, "y2": 284.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001037" }, { "image": "data/prediction/urban_night_rittenhouse/image/000330.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000330.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 414.5, "y1": 198.5, "x2": 469.0, "y2": 245.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001038" }, { "image": "data/prediction/urban_night_rittenhouse/image/000330.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000330.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.5, "y1": 198.5, "x2": 469.0, "y2": 245.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001038" }, { "image": "data/prediction/urban_night_rittenhouse/image/000331.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000331.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.0, "y1": 201.0, "x2": 519.0, "y2": 263.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001039" }, { "image": "data/prediction/urban_night_rittenhouse/image/000331.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000331.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.0, "y1": 201.0, "x2": 519.0, "y2": 263.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001039" }, { "image": "data/prediction/urban_night_rittenhouse/image/000332.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000332.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 462.0, "y1": 205.0, "x2": 604.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001040" }, { "image": "data/prediction/urban_night_rittenhouse/image/000332.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000332.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 462.0, "y1": 205.0, "x2": 604.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001040" }, { "image": "data/prediction/urban_night_rittenhouse/image/000333.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000333.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 413.0, "y1": 197.5, "x2": 466.5, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001041" }, { "image": "data/prediction/urban_night_rittenhouse/image/000333.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000333.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 413.0, "y1": 197.5, "x2": 466.5, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001041" }, { "image": "data/prediction/urban_night_rittenhouse/image/000334.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000334.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 419.0, "y1": 198.5, "x2": 498.0, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001042" }, { "image": "data/prediction/urban_night_rittenhouse/image/000334.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000334.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 419.0, "y1": 198.5, "x2": 498.0, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001042" }, { "image": "data/prediction/urban_night_rittenhouse/image/000335.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000335.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 459.5, "y1": 201.0, "x2": 578.0, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001043" }, { "image": "data/prediction/urban_night_rittenhouse/image/000335.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000335.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 459.5, "y1": 201.0, "x2": 578.0, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001043" }, { "image": "data/prediction/urban_night_rittenhouse/image/000336.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000336.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 290.0, "y1": 197.5, "x2": 342.0, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001044" }, { "image": "data/prediction/urban_night_rittenhouse/image/000336.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000336.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 290.0, "y1": 197.5, "x2": 342.0, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001044" }, { "image": "data/prediction/urban_night_rittenhouse/image/000337.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000337.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 208.5, "y1": 198.5, "x2": 298.0, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001045" }, { "image": "data/prediction/urban_night_rittenhouse/image/000337.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000337.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 208.5, "y1": 198.5, "x2": 298.0, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001045" }, { "image": "data/prediction/urban_night_rittenhouse/image/000338.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000338.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 108.0, "y1": 190.5, "x2": 212.0, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001046" }, { "image": "data/prediction/urban_night_rittenhouse/image/000338.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000338.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 108.0, "y1": 190.5, "x2": 212.0, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001046" }, { "image": "data/prediction/urban_night_rittenhouse/image/000339.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000339.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 397.0, "y1": 197.0, "x2": 456.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001047" }, { "image": "data/prediction/urban_night_rittenhouse/image/000339.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000339.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 397.0, "y1": 197.0, "x2": 456.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001047" }, { "image": "data/prediction/urban_night_rittenhouse/image/000340.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000340.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 119.0, "y1": 200.5, "x2": 224.0, "y2": 283.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001048" }, { "image": "data/prediction/urban_night_rittenhouse/image/000340.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000340.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 119.0, "y1": 200.5, "x2": 224.0, "y2": 283.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001048" }, { "image": "data/prediction/urban_night_rittenhouse/image/000341.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000341.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 260.5, "y1": 209.0, "x2": 378.0, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001049" }, { "image": "data/prediction/urban_night_rittenhouse/image/000341.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000341.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 260.5, "y1": 209.0, "x2": 378.0, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001049" }, { "image": "data/prediction/urban_night_rittenhouse/image/000342.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000342.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 401.5, "y1": 209.5, "x2": 540.5, "y2": 311.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001050" }, { "image": "data/prediction/urban_night_rittenhouse/image/000342.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000342.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 401.5, "y1": 209.5, "x2": 540.5, "y2": 311.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001050" }, { "image": "data/prediction/urban_night_rittenhouse/image/000343.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000343.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 418.5, "y1": 214.0, "x2": 591.0, "y2": 337.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001051" }, { "image": "data/prediction/urban_night_rittenhouse/image/000343.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000343.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 418.5, "y1": 214.0, "x2": 591.0, "y2": 337.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001051" }, { "image": "data/prediction/urban_night_rittenhouse/image/000344.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000344.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 431.5, "y1": 213.0, "x2": 623.0, "y2": 352.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001052" }, { "image": "data/prediction/urban_night_rittenhouse/image/000344.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000344.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 431.5, "y1": 213.0, "x2": 623.0, "y2": 352.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001052" }, { "image": "data/prediction/urban_night_rittenhouse/image/000345.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000345.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 500.5, "y1": 214.5, "x2": 637.5, "y2": 344.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001053" }, { "image": "data/prediction/urban_night_rittenhouse/image/000345.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000345.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 500.5, "y1": 214.5, "x2": 637.5, "y2": 344.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001053" }, { "image": "data/prediction/urban_night_rittenhouse/image/000346.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000346.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 16.5, "y1": 194.5, "x2": 71.5, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001054" }, { "image": "data/prediction/urban_night_rittenhouse/image/000346.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000346.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 16.5, "y1": 194.5, "x2": 71.5, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001054" }, { "image": "data/prediction/urban_night_rittenhouse/image/000347.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000347.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 168.0, "y1": 198.0, "x2": 230.0, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001055" }, { "image": "data/prediction/urban_night_rittenhouse/image/000347.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000347.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 168.0, "y1": 198.0, "x2": 230.0, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001055" }, { "image": "data/prediction/urban_night_rittenhouse/image/000348.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000348.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 311.0, "y1": 195.5, "x2": 376.0, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001056" }, { "image": "data/prediction/urban_night_rittenhouse/image/000348.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000348.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 311.0, "y1": 195.5, "x2": 376.0, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001056" }, { "image": "data/prediction/urban_night_rittenhouse/image/000349.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000349.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 327.0, "y1": 198.0, "x2": 385.0, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001057" }, { "image": "data/prediction/urban_night_rittenhouse/image/000349.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000349.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 327.0, "y1": 198.0, "x2": 385.0, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001057" }, { "image": "data/prediction/urban_night_rittenhouse/image/000350.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000350.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 324.0, "y1": 192.5, "x2": 374.5, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001058" }, { "image": "data/prediction/urban_night_rittenhouse/image/000350.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000350.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 324.0, "y1": 192.5, "x2": 374.5, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001058" }, { "image": "data/prediction/urban_night_rittenhouse/image/000351.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000351.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 403.0, "y1": 200.5, "x2": 464.0, "y2": 242.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001059" }, { "image": "data/prediction/urban_night_rittenhouse/image/000351.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000351.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.0, "y1": 200.5, "x2": 464.0, "y2": 242.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001059" }, { "image": "data/prediction/urban_night_rittenhouse/image/000352.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000352.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.5, "y1": 205.5, "x2": 474.5, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001060" }, { "image": "data/prediction/urban_night_rittenhouse/image/000352.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000352.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.5, "y1": 205.5, "x2": 474.5, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001060" }, { "image": "data/prediction/urban_night_rittenhouse/image/000353.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000353.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 418.0, "y1": 205.5, "x2": 495.0, "y2": 259.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001061" }, { "image": "data/prediction/urban_night_rittenhouse/image/000353.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000353.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 418.0, "y1": 205.5, "x2": 495.0, "y2": 259.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001061" }, { "image": "data/prediction/urban_night_rittenhouse/image/000354.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000354.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.5, "y1": 209.5, "x2": 536.0, "y2": 276.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001062" }, { "image": "data/prediction/urban_night_rittenhouse/image/000354.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000354.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.5, "y1": 209.5, "x2": 536.0, "y2": 276.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001062" }, { "image": "data/prediction/urban_night_rittenhouse/image/000355.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000355.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 476.0, "y1": 217.5, "x2": 631.5, "y2": 321.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001063" }, { "image": "data/prediction/urban_night_rittenhouse/image/000355.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000355.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 476.0, "y1": 217.5, "x2": 631.5, "y2": 321.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001063" }, { "image": "data/prediction/urban_night_rittenhouse/image/000356.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000356.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 398.5, "y1": 199.5, "x2": 452.0, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001064" }, { "image": "data/prediction/urban_night_rittenhouse/image/000356.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000356.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 398.5, "y1": 199.5, "x2": 452.0, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001064" }, { "image": "data/prediction/urban_night_rittenhouse/image/000357.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000357.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 416.0, "y1": 202.5, "x2": 491.0, "y2": 254.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001065" }, { "image": "data/prediction/urban_night_rittenhouse/image/000357.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000357.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 416.0, "y1": 202.5, "x2": 491.0, "y2": 254.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001065" }, { "image": "data/prediction/urban_night_rittenhouse/image/000358.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000358.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.5, "y1": 209.5, "x2": 571.5, "y2": 288.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001066" }, { "image": "data/prediction/urban_night_rittenhouse/image/000358.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000358.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.5, "y1": 209.5, "x2": 571.5, "y2": 288.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001066" }, { "image": "data/prediction/urban_night_rittenhouse/image/000359.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000359.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.5, "y1": 201.5, "x2": 479.0, "y2": 245.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001067" }, { "image": "data/prediction/urban_night_rittenhouse/image/000359.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000359.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.5, "y1": 201.5, "x2": 479.0, "y2": 245.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001067" }, { "image": "data/prediction/urban_night_rittenhouse/image/000360.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000360.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 451.5, "y1": 206.5, "x2": 555.5, "y2": 276.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001068" }, { "image": "data/prediction/urban_night_rittenhouse/image/000360.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000360.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 451.5, "y1": 206.5, "x2": 555.5, "y2": 276.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001068" }, { "image": "data/prediction/urban_night_rittenhouse/image/000361.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000361.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 406.0, "y1": 201.0, "x2": 482.5, "y2": 250.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001069" }, { "image": "data/prediction/urban_night_rittenhouse/image/000361.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000361.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 406.0, "y1": 201.0, "x2": 482.5, "y2": 250.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001069" }, { "image": "data/prediction/urban_night_rittenhouse/image/000362.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000362.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 446.5, "y1": 204.5, "x2": 552.0, "y2": 277.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001070" }, { "image": "data/prediction/urban_night_rittenhouse/image/000362.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000362.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 446.5, "y1": 204.5, "x2": 552.0, "y2": 277.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001070" }, { "image": "data/prediction/urban_night_rittenhouse/image/000363.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000363.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.0, "y1": 192.5, "x2": 467.0, "y2": 236.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001071" }, { "image": "data/prediction/urban_night_rittenhouse/image/000363.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000363.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.0, "y1": 192.5, "x2": 467.0, "y2": 236.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001071" }, { "image": "data/prediction/urban_night_rittenhouse/image/000364.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000364.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 434.5, "y1": 195.0, "x2": 531.0, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001072" }, { "image": "data/prediction/urban_night_rittenhouse/image/000364.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000364.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 434.5, "y1": 195.0, "x2": 531.0, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001072" }, { "image": "data/prediction/urban_night_rittenhouse/image/000365.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000365.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 403.5, "y1": 191.5, "x2": 459.0, "y2": 233.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001073" }, { "image": "data/prediction/urban_night_rittenhouse/image/000365.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000365.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 403.5, "y1": 191.5, "x2": 459.0, "y2": 233.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001073" }, { "image": "data/prediction/urban_night_rittenhouse/image/000366.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000366.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 466.5, "y1": 202.0, "x2": 635.5, "y2": 312.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001074" }, { "image": "data/prediction/urban_night_rittenhouse/image/000366.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000366.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 466.5, "y1": 202.0, "x2": 635.5, "y2": 312.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001074" }, { "image": "data/prediction/urban_night_rittenhouse/image/000367.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000367.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 404.0, "y1": 198.0, "x2": 456.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001075" }, { "image": "data/prediction/urban_night_rittenhouse/image/000367.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000367.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 404.0, "y1": 198.0, "x2": 456.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001075" }, { "image": "data/prediction/urban_night_rittenhouse/image/000368.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000368.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 423.5, "y1": 202.5, "x2": 497.5, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001076" }, { "image": "data/prediction/urban_night_rittenhouse/image/000368.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000368.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 423.5, "y1": 202.5, "x2": 497.5, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001076" }, { "image": "data/prediction/urban_night_rittenhouse/image/000369.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000369.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 455.5, "y1": 207.5, "x2": 569.0, "y2": 282.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001077" }, { "image": "data/prediction/urban_night_rittenhouse/image/000369.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000369.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 455.5, "y1": 207.5, "x2": 569.0, "y2": 282.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001077" }, { "image": "data/prediction/urban_night_rittenhouse/image/000370.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000370.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.5, "y1": 202.5, "x2": 481.0, "y2": 249.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001078" }, { "image": "data/prediction/urban_night_rittenhouse/image/000370.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000370.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.5, "y1": 202.5, "x2": 481.0, "y2": 249.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001078" }, { "image": "data/prediction/urban_night_rittenhouse/image/000371.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000371.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 443.5, "y1": 208.5, "x2": 551.5, "y2": 278.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001079" }, { "image": "data/prediction/urban_night_rittenhouse/image/000371.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000371.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 443.5, "y1": 208.5, "x2": 551.5, "y2": 278.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001079" }, { "image": "data/prediction/urban_night_rittenhouse/image/000372.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000372.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 338.0, "y1": 196.0, "x2": 388.5, "y2": 240.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001080" }, { "image": "data/prediction/urban_night_rittenhouse/image/000372.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000372.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 338.0, "y1": 196.0, "x2": 388.5, "y2": 240.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001080" }, { "image": "data/prediction/urban_night_rittenhouse/image/000373.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000373.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 344.0, "y1": 198.0, "x2": 400.0, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001081" }, { "image": "data/prediction/urban_night_rittenhouse/image/000373.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000373.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 344.0, "y1": 198.0, "x2": 400.0, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001081" }, { "image": "data/prediction/urban_night_rittenhouse/image/000374.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000374.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 331.0, "y1": 200.5, "x2": 407.5, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001082" }, { "image": "data/prediction/urban_night_rittenhouse/image/000374.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000374.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.0, "y1": 200.5, "x2": 407.5, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001082" }, { "image": "data/prediction/urban_night_rittenhouse/image/000375.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000375.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 291.5, "y1": 201.5, "x2": 401.5, "y2": 271.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001083" }, { "image": "data/prediction/urban_night_rittenhouse/image/000375.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000375.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 291.5, "y1": 201.5, "x2": 401.5, "y2": 271.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001083" }, { "image": "data/prediction/urban_night_rittenhouse/image/000376.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000376.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 202.5, "y1": 199.5, "x2": 361.0, "y2": 287.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001084" }, { "image": "data/prediction/urban_night_rittenhouse/image/000376.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000376.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 202.5, "y1": 199.5, "x2": 361.0, "y2": 287.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001084" }, { "image": "data/prediction/urban_night_rittenhouse/image/000377.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000377.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 55.0, "y1": 201.5, "x2": 255.5, "y2": 314.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001085" }, { "image": "data/prediction/urban_night_rittenhouse/image/000377.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000377.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 55.0, "y1": 201.5, "x2": 255.5, "y2": 314.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001085" }, { "image": "data/prediction/urban_night_rittenhouse/image/000378.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000378.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 196.0, "x2": 26.0, "y2": 259.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001086" }, { "image": "data/prediction/urban_night_rittenhouse/image/000378.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000378.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 196.0, "x2": 26.0, "y2": 259.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001086" }, { "image": "data/prediction/urban_night_rittenhouse/image/000379.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000379.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 116.0, "y1": 198.0, "x2": 253.5, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001087" }, { "image": "data/prediction/urban_night_rittenhouse/image/000379.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000379.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 116.0, "y1": 198.0, "x2": 253.5, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001087" }, { "image": "data/prediction/urban_night_rittenhouse/image/000380.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000380.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 369.5, "y1": 204.0, "x2": 563.0, "y2": 324.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001088" }, { "image": "data/prediction/urban_night_rittenhouse/image/000380.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000380.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 369.5, "y1": 204.0, "x2": 563.0, "y2": 324.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001088" }, { "image": "data/prediction/urban_night_rittenhouse/image/000381.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000381.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 518.5, "y1": 217.0, "x2": 639.5, "y2": 389.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001089" }, { "image": "data/prediction/urban_night_rittenhouse/image/000381.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000381.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 518.5, "y1": 217.0, "x2": 639.5, "y2": 389.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001089" }, { "image": "data/prediction/urban_night_rittenhouse/image/000382.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000382.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 225.5, "x2": 117.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001090" }, { "image": "data/prediction/urban_night_rittenhouse/image/000382.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000382.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 225.5, "x2": 117.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001090" }, { "image": "data/prediction/urban_night_rittenhouse/image/000383.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000383.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 33.0, "y1": 224.5, "x2": 226.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001091" }, { "image": "data/prediction/urban_night_rittenhouse/image/000383.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000383.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 33.0, "y1": 224.5, "x2": 226.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001091" }, { "image": "data/prediction/urban_night_rittenhouse/image/000384.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000384.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 209.5, "y1": 212.5, "x2": 400.0, "y2": 383.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001092" }, { "image": "data/prediction/urban_night_rittenhouse/image/000384.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000384.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 209.5, "y1": 212.5, "x2": 400.0, "y2": 383.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001092" }, { "image": "data/prediction/urban_night_rittenhouse/image/000385.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000385.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 255.0, "y1": 213.0, "x2": 420.5, "y2": 363.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001093" }, { "image": "data/prediction/urban_night_rittenhouse/image/000385.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000385.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 255.0, "y1": 213.0, "x2": 420.5, "y2": 363.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001093" }, { "image": "data/prediction/urban_night_rittenhouse/image/000386.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000386.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 264.5, "y1": 214.5, "x2": 410.0, "y2": 348.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001094" }, { "image": "data/prediction/urban_night_rittenhouse/image/000386.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000386.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 264.5, "y1": 214.5, "x2": 410.0, "y2": 348.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001094" }, { "image": "data/prediction/urban_night_rittenhouse/image/000387.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000387.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 263.5, "y1": 215.5, "x2": 401.5, "y2": 343.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001095" }, { "image": "data/prediction/urban_night_rittenhouse/image/000387.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000387.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 263.5, "y1": 215.5, "x2": 401.5, "y2": 343.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001095" }, { "image": "data/prediction/urban_night_rittenhouse/image/000388.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000388.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 268.5, "y1": 210.5, "x2": 402.0, "y2": 331.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001096" }, { "image": "data/prediction/urban_night_rittenhouse/image/000388.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000388.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 268.5, "y1": 210.5, "x2": 402.0, "y2": 331.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001096" }, { "image": "data/prediction/urban_night_rittenhouse/image/000389.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000389.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 267.5, "y1": 209.0, "x2": 397.0, "y2": 325.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001097" }, { "image": "data/prediction/urban_night_rittenhouse/image/000389.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000389.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 267.5, "y1": 209.0, "x2": 397.0, "y2": 325.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001097" }, { "image": "data/prediction/urban_night_rittenhouse/image/000390.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000390.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 188.5, "y1": 198.0, "x2": 285.5, "y2": 253.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001098" }, { "image": "data/prediction/urban_night_rittenhouse/image/000390.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000390.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 188.5, "y1": 198.0, "x2": 285.5, "y2": 253.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001098" }, { "image": "data/prediction/urban_night_rittenhouse/image/000391.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000391.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 401.5, "y1": 200.0, "x2": 524.0, "y2": 270.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001099" }, { "image": "data/prediction/urban_night_rittenhouse/image/000391.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000391.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 401.5, "y1": 200.0, "x2": 524.0, "y2": 270.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001099" }, { "image": "data/prediction/urban_night_rittenhouse/image/000392.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000392.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 459.5, "y1": 212.5, "x2": 623.0, "y2": 308.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001100" }, { "image": "data/prediction/urban_night_rittenhouse/image/000392.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000392.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 459.5, "y1": 212.5, "x2": 623.0, "y2": 308.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001100" }, { "image": "data/prediction/urban_night_rittenhouse/image/000393.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000393.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 510.5, "y1": 221.5, "x2": 639.5, "y2": 365.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001101" }, { "image": "data/prediction/urban_night_rittenhouse/image/000393.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000393.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 510.5, "y1": 221.5, "x2": 639.5, "y2": 365.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001101" }, { "image": "data/prediction/urban_night_rittenhouse/image/000394.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000394.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 258.5, "y1": 188.5, "x2": 329.0, "y2": 216.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001102" }, { "image": "data/prediction/urban_night_rittenhouse/image/000394.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000394.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 258.5, "y1": 188.5, "x2": 329.0, "y2": 216.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001102" }, { "image": "data/prediction/urban_night_rittenhouse/image/000395.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000395.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 505.0, "y1": 209.0, "x2": 639.5, "y2": 284.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001103" }, { "image": "data/prediction/urban_night_rittenhouse/image/000395.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000395.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 505.0, "y1": 209.0, "x2": 639.5, "y2": 284.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001103" }, { "image": "data/prediction/urban_night_rittenhouse/image/000396.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000396.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 265.0, "y1": 211.0, "x2": 385.0, "y2": 320.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001104" }, { "image": "data/prediction/urban_night_rittenhouse/image/000396.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000396.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 265.0, "y1": 211.0, "x2": 385.0, "y2": 320.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001104" }, { "image": "data/prediction/urban_night_rittenhouse/image/000397.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000397.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 266.0, "y1": 204.0, "x2": 372.5, "y2": 298.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001105" }, { "image": "data/prediction/urban_night_rittenhouse/image/000397.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000397.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 266.0, "y1": 204.0, "x2": 372.5, "y2": 298.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001105" }, { "image": "data/prediction/urban_night_rittenhouse/image/000398.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000398.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 277.0, "y1": 199.0, "x2": 373.5, "y2": 284.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001106" }, { "image": "data/prediction/urban_night_rittenhouse/image/000398.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000398.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 277.0, "y1": 199.0, "x2": 373.5, "y2": 284.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001106" }, { "image": "data/prediction/urban_night_rittenhouse/image/000399.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000399.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 287.0, "y1": 202.5, "x2": 383.0, "y2": 287.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001107" }, { "image": "data/prediction/urban_night_rittenhouse/image/000399.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000399.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 287.0, "y1": 202.5, "x2": 383.0, "y2": 287.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001107" }, { "image": "data/prediction/urban_night_rittenhouse/image/000400.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000400.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 282.5, "y1": 211.0, "x2": 377.5, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001108" }, { "image": "data/prediction/urban_night_rittenhouse/image/000400.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000400.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 282.5, "y1": 211.0, "x2": 377.5, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001108" }, { "image": "data/prediction/urban_night_rittenhouse/image/000401.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000401.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 280.5, "y1": 209.5, "x2": 377.0, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001109" }, { "image": "data/prediction/urban_night_rittenhouse/image/000401.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000401.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.5, "y1": 209.5, "x2": 377.0, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001109" }, { "image": "data/prediction/urban_night_rittenhouse/image/000402.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000402.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 268.0, "y1": 208.0, "x2": 368.5, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001110" }, { "image": "data/prediction/urban_night_rittenhouse/image/000402.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000402.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 268.0, "y1": 208.0, "x2": 368.5, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001110" }, { "image": "data/prediction/urban_night_rittenhouse/image/000403.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000403.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 267.0, "y1": 206.0, "x2": 370.5, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001111" }, { "image": "data/prediction/urban_night_rittenhouse/image/000403.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000403.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 267.0, "y1": 206.0, "x2": 370.5, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001111" }, { "image": "data/prediction/urban_night_rittenhouse/image/000404.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000404.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 267.0, "y1": 205.0, "x2": 375.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001112" }, { "image": "data/prediction/urban_night_rittenhouse/image/000404.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000404.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 267.0, "y1": 205.0, "x2": 375.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001112" }, { "image": "data/prediction/urban_night_rittenhouse/image/000405.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000405.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 280.5, "y1": 206.0, "x2": 391.0, "y2": 303.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001113" }, { "image": "data/prediction/urban_night_rittenhouse/image/000405.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000405.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.5, "y1": 206.0, "x2": 391.0, "y2": 303.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001113" }, { "image": "data/prediction/urban_night_rittenhouse/image/000406.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000406.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 384.0, "y1": 190.5, "x2": 465.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001114" }, { "image": "data/prediction/urban_night_rittenhouse/image/000406.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000406.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 384.0, "y1": 190.5, "x2": 465.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001114" }, { "image": "data/prediction/urban_night_rittenhouse/image/000407.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000407.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 389.5, "y1": 194.0, "x2": 491.0, "y2": 249.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001115" }, { "image": "data/prediction/urban_night_rittenhouse/image/000407.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000407.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 389.5, "y1": 194.0, "x2": 491.0, "y2": 249.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001115" }, { "image": "data/prediction/urban_night_rittenhouse/image/000408.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000408.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 409.0, "y1": 193.0, "x2": 532.5, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001116" }, { "image": "data/prediction/urban_night_rittenhouse/image/000408.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000408.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 409.0, "y1": 193.0, "x2": 532.5, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001116" }, { "image": "data/prediction/urban_night_rittenhouse/image/000409.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000409.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 444.0, "y1": 192.5, "x2": 598.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001117" }, { "image": "data/prediction/urban_night_rittenhouse/image/000409.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000409.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 444.0, "y1": 192.5, "x2": 598.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001117" }, { "image": "data/prediction/urban_night_rittenhouse/image/000410.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000410.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 280.0, "y1": 205.5, "x2": 390.0, "y2": 303.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001118" }, { "image": "data/prediction/urban_night_rittenhouse/image/000410.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000410.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.0, "y1": 205.5, "x2": 390.0, "y2": 303.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001118" }, { "image": "data/prediction/urban_night_rittenhouse/image/000411.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000411.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 272.5, "y1": 206.5, "x2": 377.0, "y2": 300.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001119" }, { "image": "data/prediction/urban_night_rittenhouse/image/000411.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000411.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 272.5, "y1": 206.5, "x2": 377.0, "y2": 300.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001119" }, { "image": "data/prediction/urban_night_rittenhouse/image/000412.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000412.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 272.5, "y1": 207.0, "x2": 370.5, "y2": 295.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001120" }, { "image": "data/prediction/urban_night_rittenhouse/image/000412.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000412.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 272.5, "y1": 207.0, "x2": 370.5, "y2": 295.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001120" }, { "image": "data/prediction/urban_night_rittenhouse/image/000413.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000413.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 279.5, "y1": 207.0, "x2": 379.0, "y2": 295.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001121" }, { "image": "data/prediction/urban_night_rittenhouse/image/000413.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000413.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 279.5, "y1": 207.0, "x2": 379.0, "y2": 295.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001121" }, { "image": "data/prediction/urban_night_rittenhouse/image/000414.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000414.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 285.0, "y1": 205.0, "x2": 390.5, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001122" }, { "image": "data/prediction/urban_night_rittenhouse/image/000414.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000414.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 285.0, "y1": 205.0, "x2": 390.5, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001122" }, { "image": "data/prediction/urban_night_rittenhouse/image/000415.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000415.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 290.0, "y1": 206.5, "x2": 399.5, "y2": 304.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001123" }, { "image": "data/prediction/urban_night_rittenhouse/image/000415.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000415.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 290.0, "y1": 206.5, "x2": 399.5, "y2": 304.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001123" }, { "image": "data/prediction/urban_night_rittenhouse/image/000416.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000416.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 284.5, "y1": 211.0, "x2": 402.0, "y2": 318.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001124" }, { "image": "data/prediction/urban_night_rittenhouse/image/000416.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000416.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 284.5, "y1": 211.0, "x2": 402.0, "y2": 318.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001124" }, { "image": "data/prediction/urban_night_rittenhouse/image/000417.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000417.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 280.0, "y1": 211.0, "x2": 405.5, "y2": 324.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001125" }, { "image": "data/prediction/urban_night_rittenhouse/image/000417.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000417.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.0, "y1": 211.0, "x2": 405.5, "y2": 324.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001125" }, { "image": "data/prediction/urban_night_rittenhouse/image/000418.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000418.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 280.5, "y1": 206.0, "x2": 406.5, "y2": 317.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001126" }, { "image": "data/prediction/urban_night_rittenhouse/image/000418.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000418.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 280.5, "y1": 206.0, "x2": 406.5, "y2": 317.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001126" }, { "image": "data/prediction/urban_night_rittenhouse/image/000419.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000419.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 284.0, "y1": 208.5, "x2": 407.0, "y2": 318.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001127" }, { "image": "data/prediction/urban_night_rittenhouse/image/000419.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000419.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 284.0, "y1": 208.5, "x2": 407.0, "y2": 318.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001127" }, { "image": "data/prediction/urban_night_rittenhouse/image/000420.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000420.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 294.5, "y1": 206.0, "x2": 407.5, "y2": 307.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001128" }, { "image": "data/prediction/urban_night_rittenhouse/image/000420.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000420.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 294.5, "y1": 206.0, "x2": 407.5, "y2": 307.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001128" }, { "image": "data/prediction/urban_night_rittenhouse/image/000421.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000421.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 299.5, "y1": 210.0, "x2": 408.0, "y2": 309.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001129" }, { "image": "data/prediction/urban_night_rittenhouse/image/000421.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000421.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.5, "y1": 210.0, "x2": 408.0, "y2": 309.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001129" }, { "image": "data/prediction/urban_night_rittenhouse/image/000422.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000422.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 298.5, "y1": 210.0, "x2": 407.0, "y2": 308.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001130" }, { "image": "data/prediction/urban_night_rittenhouse/image/000422.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000422.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 298.5, "y1": 210.0, "x2": 407.0, "y2": 308.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001130" }, { "image": "data/prediction/urban_night_rittenhouse/image/000423.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000423.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 300.0, "y1": 207.5, "x2": 410.0, "y2": 306.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001131" }, { "image": "data/prediction/urban_night_rittenhouse/image/000423.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000423.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 300.0, "y1": 207.5, "x2": 410.0, "y2": 306.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001131" }, { "image": "data/prediction/urban_night_rittenhouse/image/000424.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000424.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 299.0, "y1": 205.5, "x2": 406.0, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001132" }, { "image": "data/prediction/urban_night_rittenhouse/image/000424.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000424.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.0, "y1": 205.5, "x2": 406.0, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001132" }, { "image": "data/prediction/urban_night_rittenhouse/image/000425.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000425.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 297.5, "y1": 208.5, "x2": 402.5, "y2": 304.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001133" }, { "image": "data/prediction/urban_night_rittenhouse/image/000425.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000425.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 297.5, "y1": 208.5, "x2": 402.5, "y2": 304.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001133" }, { "image": "data/prediction/urban_night_rittenhouse/image/000426.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000426.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 303.5, "y1": 206.0, "x2": 409.5, "y2": 302.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001134" }, { "image": "data/prediction/urban_night_rittenhouse/image/000426.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000426.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 303.5, "y1": 206.0, "x2": 409.5, "y2": 302.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001134" }, { "image": "data/prediction/urban_night_rittenhouse/image/000427.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000427.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 297.0, "y1": 208.0, "x2": 414.0, "y2": 314.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001135" }, { "image": "data/prediction/urban_night_rittenhouse/image/000427.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000427.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 297.0, "y1": 208.0, "x2": 414.0, "y2": 314.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001135" }, { "image": "data/prediction/urban_night_rittenhouse/image/000428.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000428.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 274.0, "y1": 205.0, "x2": 420.5, "y2": 335.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001136" }, { "image": "data/prediction/urban_night_rittenhouse/image/000428.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000428.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 274.0, "y1": 205.0, "x2": 420.5, "y2": 335.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001136" }, { "image": "data/prediction/urban_night_rittenhouse/image/000429.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000429.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 245.5, "y1": 208.5, "x2": 427.5, "y2": 372.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001137" }, { "image": "data/prediction/urban_night_rittenhouse/image/000429.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000429.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 245.5, "y1": 208.5, "x2": 427.5, "y2": 372.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001137" }, { "image": "data/prediction/urban_night_rittenhouse/image/000430.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000430.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 441.0, "y1": 204.0, "x2": 515.0, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001138" }, { "image": "data/prediction/urban_night_rittenhouse/image/000430.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000430.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.0, "y1": 204.0, "x2": 515.0, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001138" }, { "image": "data/prediction/urban_night_rittenhouse/image/000431.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000431.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.5, "y1": 204.0, "x2": 548.0, "y2": 269.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001139" }, { "image": "data/prediction/urban_night_rittenhouse/image/000431.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000431.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.5, "y1": 204.0, "x2": 548.0, "y2": 269.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001139" }, { "image": "data/prediction/urban_night_rittenhouse/image/000432.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000432.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 473.0, "y1": 209.5, "x2": 602.5, "y2": 294.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001140" }, { "image": "data/prediction/urban_night_rittenhouse/image/000432.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000432.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 473.0, "y1": 209.5, "x2": 602.5, "y2": 294.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001140" }, { "image": "data/prediction/urban_night_rittenhouse/image/000433.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000433.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 516.5, "y1": 216.5, "x2": 639.5, "y2": 342.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001141" }, { "image": "data/prediction/urban_night_rittenhouse/image/000433.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000433.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 516.5, "y1": 216.5, "x2": 639.5, "y2": 342.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001141" }, { "image": "data/prediction/urban_night_rittenhouse/image/000434.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000434.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 417.0, "y1": 193.0, "x2": 477.0, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001142" }, { "image": "data/prediction/urban_night_rittenhouse/image/000434.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000434.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 417.0, "y1": 193.0, "x2": 477.0, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001142" }, { "image": "data/prediction/urban_night_rittenhouse/image/000435.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000435.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 426.0, "y1": 195.0, "x2": 503.5, "y2": 250.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001143" }, { "image": "data/prediction/urban_night_rittenhouse/image/000435.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000435.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 426.0, "y1": 195.0, "x2": 503.5, "y2": 250.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001143" }, { "image": "data/prediction/urban_night_rittenhouse/image/000436.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000436.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 441.5, "y1": 196.0, "x2": 544.5, "y2": 268.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001144" }, { "image": "data/prediction/urban_night_rittenhouse/image/000436.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000436.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.5, "y1": 196.0, "x2": 544.5, "y2": 268.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001144" }, { "image": "data/prediction/urban_night_rittenhouse/image/000437.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000437.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 469.5, "y1": 197.5, "x2": 628.5, "y2": 301.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001145" }, { "image": "data/prediction/urban_night_rittenhouse/image/000437.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000437.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 469.5, "y1": 197.5, "x2": 628.5, "y2": 301.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001145" }, { "image": "data/prediction/urban_night_rittenhouse/image/000438.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000438.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 512.0, "y1": 200.0, "x2": 639.5, "y2": 358.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001146" }, { "image": "data/prediction/urban_night_rittenhouse/image/000438.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000438.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 512.0, "y1": 200.0, "x2": 639.5, "y2": 358.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001146" }, { "image": "data/prediction/urban_night_rittenhouse/image/000439.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000439.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 234.5, "y1": 212.0, "x2": 441.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001147" }, { "image": "data/prediction/urban_night_rittenhouse/image/000439.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000439.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 234.5, "y1": 212.0, "x2": 441.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001147" }, { "image": "data/prediction/urban_night_rittenhouse/image/000440.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000440.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 222.0, "y1": 214.5, "x2": 430.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001148" }, { "image": "data/prediction/urban_night_rittenhouse/image/000440.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000440.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 222.0, "y1": 214.5, "x2": 430.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001148" }, { "image": "data/prediction/urban_night_rittenhouse/image/000441.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000441.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 185.0, "y1": 216.0, "x2": 410.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001149" }, { "image": "data/prediction/urban_night_rittenhouse/image/000441.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000441.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 185.0, "y1": 216.0, "x2": 410.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001149" }, { "image": "data/prediction/urban_night_rittenhouse/image/000442.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000442.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 92.0, "y1": 216.5, "x2": 331.0, "y2": 381.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001150" }, { "image": "data/prediction/urban_night_rittenhouse/image/000442.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000442.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 92.0, "y1": 216.5, "x2": 331.0, "y2": 381.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001150" }, { "image": "data/prediction/urban_night_rittenhouse/image/000443.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000443.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 215.5, "x2": 186.5, "y2": 342.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001151" }, { "image": "data/prediction/urban_night_rittenhouse/image/000443.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000443.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 215.5, "x2": 186.5, "y2": 342.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001151" }, { "image": "data/prediction/urban_night_rittenhouse/image/000444.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000444.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 208.0, "x2": 97.5, "y2": 312.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001152" }, { "image": "data/prediction/urban_night_rittenhouse/image/000444.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000444.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 208.0, "x2": 97.5, "y2": 312.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001152" }, { "image": "data/prediction/urban_night_rittenhouse/image/000445.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000445.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 0.0, "y1": 206.5, "x2": 81.0, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001153" }, { "image": "data/prediction/urban_night_rittenhouse/image/000445.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000445.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 206.5, "x2": 81.0, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001153" }, { "image": "data/prediction/urban_night_rittenhouse/image/000446.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000446.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 204.0, "x2": 33.5, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001154" }, { "image": "data/prediction/urban_night_rittenhouse/image/000446.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000446.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 204.0, "x2": 33.5, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001154" }, { "image": "data/prediction/urban_night_rittenhouse/image/000447.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000447.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 3.5, "y1": 217.5, "x2": 168.0, "y2": 338.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001155" }, { "image": "data/prediction/urban_night_rittenhouse/image/000447.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000447.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 3.5, "y1": 217.5, "x2": 168.0, "y2": 338.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001155" }, { "image": "data/prediction/urban_night_rittenhouse/image/000448.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000448.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 28.0, "y1": 201.0, "x2": 88.0, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001156" }, { "image": "data/prediction/urban_night_rittenhouse/image/000448.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000448.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 28.0, "y1": 201.0, "x2": 88.0, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001156" }, { "image": "data/prediction/urban_night_rittenhouse/image/000449.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000449.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 107.5, "y1": 209.5, "x2": 202.5, "y2": 288.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001157" }, { "image": "data/prediction/urban_night_rittenhouse/image/000449.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000449.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 107.5, "y1": 209.5, "x2": 202.5, "y2": 288.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001157" }, { "image": "data/prediction/urban_night_rittenhouse/image/000450.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000450.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 35.0, "y1": 219.0, "x2": 230.0, "y2": 361.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001158" }, { "image": "data/prediction/urban_night_rittenhouse/image/000450.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000450.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 35.0, "y1": 219.0, "x2": 230.0, "y2": 361.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001158" }, { "image": "data/prediction/urban_night_rittenhouse/image/000451.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000451.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 216.0, "x2": 22.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001159" }, { "image": "data/prediction/urban_night_rittenhouse/image/000451.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000451.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 216.0, "x2": 22.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001159" }, { "image": "data/prediction/urban_night_rittenhouse/image/000452.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000452.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 173.5, "y1": 199.0, "x2": 231.0, "y2": 234.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001160" }, { "image": "data/prediction/urban_night_rittenhouse/image/000452.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000452.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 173.5, "y1": 199.0, "x2": 231.0, "y2": 234.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001160" }, { "image": "data/prediction/urban_night_rittenhouse/image/000453.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000453.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 226.0, "y1": 199.0, "x2": 285.0, "y2": 243.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001161" }, { "image": "data/prediction/urban_night_rittenhouse/image/000453.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000453.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 226.0, "y1": 199.0, "x2": 285.0, "y2": 243.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001161" }, { "image": "data/prediction/urban_night_rittenhouse/image/000454.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000454.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 201.0, "y1": 199.5, "x2": 279.5, "y2": 265.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001162" }, { "image": "data/prediction/urban_night_rittenhouse/image/000454.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000454.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 201.0, "y1": 199.5, "x2": 279.5, "y2": 265.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001162" }, { "image": "data/prediction/urban_night_rittenhouse/image/000455.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000455.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 84.0, "y1": 202.5, "x2": 246.5, "y2": 327.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001163" }, { "image": "data/prediction/urban_night_rittenhouse/image/000455.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000455.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 84.0, "y1": 202.5, "x2": 246.5, "y2": 327.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001163" }, { "image": "data/prediction/urban_night_rittenhouse/image/000456.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000456.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 165.0, "y1": 199.0, "x2": 227.0, "y2": 244.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001164" }, { "image": "data/prediction/urban_night_rittenhouse/image/000456.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000456.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 165.0, "y1": 199.0, "x2": 227.0, "y2": 244.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001164" }, { "image": "data/prediction/urban_night_rittenhouse/image/000457.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000457.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 188.0, "y1": 201.5, "x2": 268.0, "y2": 261.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001165" }, { "image": "data/prediction/urban_night_rittenhouse/image/000457.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000457.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 188.0, "y1": 201.5, "x2": 268.0, "y2": 261.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001165" }, { "image": "data/prediction/urban_night_rittenhouse/image/000458.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000458.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 112.0, "y1": 205.5, "x2": 259.5, "y2": 309.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001166" }, { "image": "data/prediction/urban_night_rittenhouse/image/000458.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000458.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 112.0, "y1": 205.5, "x2": 259.5, "y2": 309.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001166" }, { "image": "data/prediction/urban_night_rittenhouse/image/000459.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000459.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 248.5, "y1": 196.5, "x2": 308.0, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001167" }, { "image": "data/prediction/urban_night_rittenhouse/image/000459.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000459.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 248.5, "y1": 196.5, "x2": 308.0, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001167" }, { "image": "data/prediction/urban_night_rittenhouse/image/000460.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000460.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 235.5, "y1": 195.5, "x2": 311.5, "y2": 247.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001168" }, { "image": "data/prediction/urban_night_rittenhouse/image/000460.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000460.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 235.5, "y1": 195.5, "x2": 311.5, "y2": 247.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001168" }, { "image": "data/prediction/urban_night_rittenhouse/image/000461.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000461.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 170.0, "y1": 202.0, "x2": 284.0, "y2": 286.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001169" }, { "image": "data/prediction/urban_night_rittenhouse/image/000461.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000461.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 170.0, "y1": 202.0, "x2": 284.0, "y2": 286.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001169" }, { "image": "data/prediction/urban_night_rittenhouse/image/000462.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000462.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 145.5, "y1": 200.5, "x2": 240.0, "y2": 271.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001170" }, { "image": "data/prediction/urban_night_rittenhouse/image/000462.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000462.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 145.5, "y1": 200.5, "x2": 240.0, "y2": 271.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001170" }, { "image": "data/prediction/urban_night_rittenhouse/image/000463.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000463.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 51.0, "y1": 209.0, "x2": 232.5, "y2": 336.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001171" }, { "image": "data/prediction/urban_night_rittenhouse/image/000463.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000463.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 51.0, "y1": 209.0, "x2": 232.5, "y2": 336.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001171" }, { "image": "data/prediction/urban_night_rittenhouse/image/000464.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000464.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 218.5, "y1": 196.5, "x2": 291.5, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001172" }, { "image": "data/prediction/urban_night_rittenhouse/image/000464.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000464.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 218.5, "y1": 196.5, "x2": 291.5, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001172" }, { "image": "data/prediction/urban_night_rittenhouse/image/000465.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000465.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 231.0, "y1": 190.5, "x2": 292.5, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001173" }, { "image": "data/prediction/urban_night_rittenhouse/image/000465.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000465.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 231.0, "y1": 190.5, "x2": 292.5, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001173" }, { "image": "data/prediction/urban_night_rittenhouse/image/000466.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000466.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 186.5, "y1": 195.0, "x2": 275.0, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001174" }, { "image": "data/prediction/urban_night_rittenhouse/image/000466.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000466.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 186.5, "y1": 195.0, "x2": 275.0, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001174" }, { "image": "data/prediction/urban_night_rittenhouse/image/000467.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000467.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 92.5, "y1": 200.5, "x2": 247.5, "y2": 316.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001175" }, { "image": "data/prediction/urban_night_rittenhouse/image/000467.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000467.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 92.5, "y1": 200.5, "x2": 247.5, "y2": 316.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001175" }, { "image": "data/prediction/urban_night_rittenhouse/image/000468.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000468.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 237.5, "y1": 197.5, "x2": 295.5, "y2": 244.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001176" }, { "image": "data/prediction/urban_night_rittenhouse/image/000468.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000468.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 237.5, "y1": 197.5, "x2": 295.5, "y2": 244.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001176" }, { "image": "data/prediction/urban_night_rittenhouse/image/000469.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000469.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 197.0, "y1": 198.0, "x2": 283.0, "y2": 267.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001177" }, { "image": "data/prediction/urban_night_rittenhouse/image/000469.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000469.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 197.0, "y1": 198.0, "x2": 283.0, "y2": 267.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001177" }, { "image": "data/prediction/urban_night_rittenhouse/image/000470.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000470.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 100.0, "y1": 204.0, "x2": 259.5, "y2": 327.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001178" }, { "image": "data/prediction/urban_night_rittenhouse/image/000470.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000470.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 100.0, "y1": 204.0, "x2": 259.5, "y2": 327.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001178" }, { "image": "data/prediction/urban_night_rittenhouse/image/000471.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000471.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 112.0, "y1": 207.5, "x2": 256.5, "y2": 317.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001179" }, { "image": "data/prediction/urban_night_rittenhouse/image/000471.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000471.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 112.0, "y1": 207.5, "x2": 256.5, "y2": 317.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001179" }, { "image": "data/prediction/urban_night_rittenhouse/image/000472.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000472.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 243.0, "y1": 200.0, "x2": 293.5, "y2": 245.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001180" }, { "image": "data/prediction/urban_night_rittenhouse/image/000472.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000472.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.0, "y1": 200.0, "x2": 293.5, "y2": 245.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001180" }, { "image": "data/prediction/urban_night_rittenhouse/image/000473.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000473.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 201.5, "y1": 200.0, "x2": 277.0, "y2": 261.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001181" }, { "image": "data/prediction/urban_night_rittenhouse/image/000473.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000473.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 201.5, "y1": 200.0, "x2": 277.0, "y2": 261.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001181" }, { "image": "data/prediction/urban_night_rittenhouse/image/000474.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000474.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 117.0, "y1": 208.5, "x2": 256.5, "y2": 308.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001182" }, { "image": "data/prediction/urban_night_rittenhouse/image/000474.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000474.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 117.0, "y1": 208.5, "x2": 256.5, "y2": 308.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001182" }, { "image": "data/prediction/urban_night_rittenhouse/image/000475.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000475.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 243.5, "y1": 192.5, "x2": 294.5, "y2": 237.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001183" }, { "image": "data/prediction/urban_night_rittenhouse/image/000475.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000475.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.5, "y1": 192.5, "x2": 294.5, "y2": 237.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001183" }, { "image": "data/prediction/urban_night_rittenhouse/image/000476.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000476.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 205.5, "y1": 190.5, "x2": 285.0, "y2": 256.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001184" }, { "image": "data/prediction/urban_night_rittenhouse/image/000476.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000476.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 205.5, "y1": 190.5, "x2": 285.0, "y2": 256.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001184" }, { "image": "data/prediction/urban_night_rittenhouse/image/000477.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000477.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 128.0, "y1": 194.5, "x2": 265.0, "y2": 301.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001185" }, { "image": "data/prediction/urban_night_rittenhouse/image/000477.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000477.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 128.0, "y1": 194.5, "x2": 265.0, "y2": 301.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001185" }, { "image": "data/prediction/urban_night_rittenhouse/image/000478.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000478.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 203.5, "x2": 222.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001186" }, { "image": "data/prediction/urban_night_rittenhouse/image/000478.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000478.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 203.5, "x2": 222.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001186" }, { "image": "data/prediction/urban_night_rittenhouse/image/000479.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000479.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 206.0, "y1": 204.5, "x2": 287.0, "y2": 263.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001187" }, { "image": "data/prediction/urban_night_rittenhouse/image/000479.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000479.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 206.0, "y1": 204.5, "x2": 287.0, "y2": 263.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001187" }, { "image": "data/prediction/urban_night_rittenhouse/image/000480.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000480.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 138.0, "y1": 210.5, "x2": 260.5, "y2": 303.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001188" }, { "image": "data/prediction/urban_night_rittenhouse/image/000480.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000480.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 138.0, "y1": 210.5, "x2": 260.5, "y2": 303.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001188" }, { "image": "data/prediction/urban_night_rittenhouse/image/000481.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000481.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 241.5, "y1": 196.0, "x2": 312.0, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001189" }, { "image": "data/prediction/urban_night_rittenhouse/image/000481.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000481.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 241.5, "y1": 196.0, "x2": 312.0, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001189" }, { "image": "data/prediction/urban_night_rittenhouse/image/000482.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000482.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 207.0, "y1": 196.5, "x2": 294.0, "y2": 256.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001190" }, { "image": "data/prediction/urban_night_rittenhouse/image/000482.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000482.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 207.0, "y1": 196.5, "x2": 294.0, "y2": 256.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001190" }, { "image": "data/prediction/urban_night_rittenhouse/image/000483.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000483.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 138.5, "y1": 199.0, "x2": 267.0, "y2": 293.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001191" }, { "image": "data/prediction/urban_night_rittenhouse/image/000483.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000483.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 138.5, "y1": 199.0, "x2": 267.0, "y2": 293.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001191" }, { "image": "data/prediction/urban_night_rittenhouse/image/000484.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000484.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 216.0, "x2": 214.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001192" }, { "image": "data/prediction/urban_night_rittenhouse/image/000484.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000484.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 216.0, "x2": 214.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001192" }, { "image": "data/prediction/urban_night_rittenhouse/image/000485.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000485.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 265.0, "y1": 193.5, "x2": 319.5, "y2": 226.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001193" }, { "image": "data/prediction/urban_night_rittenhouse/image/000485.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000485.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 265.0, "y1": 193.5, "x2": 319.5, "y2": 226.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001193" }, { "image": "data/prediction/urban_night_rittenhouse/image/000486.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000486.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 219.0, "y1": 195.0, "x2": 280.0, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001194" }, { "image": "data/prediction/urban_night_rittenhouse/image/000486.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000486.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 219.0, "y1": 195.0, "x2": 280.0, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001194" }, { "image": "data/prediction/urban_night_rittenhouse/image/000487.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000487.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 179.5, "y1": 199.5, "x2": 265.5, "y2": 273.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001195" }, { "image": "data/prediction/urban_night_rittenhouse/image/000487.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000487.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 179.5, "y1": 199.5, "x2": 265.5, "y2": 273.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001195" }, { "image": "data/prediction/urban_night_rittenhouse/image/000488.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000488.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 118.0, "y1": 202.5, "x2": 257.0, "y2": 308.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001196" }, { "image": "data/prediction/urban_night_rittenhouse/image/000488.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000488.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 118.0, "y1": 202.5, "x2": 257.0, "y2": 308.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001196" }, { "image": "data/prediction/urban_night_rittenhouse/image/000489.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000489.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 243.0, "y1": 196.5, "x2": 302.5, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001197" }, { "image": "data/prediction/urban_night_rittenhouse/image/000489.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000489.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.0, "y1": 196.5, "x2": 302.5, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001197" }, { "image": "data/prediction/urban_night_rittenhouse/image/000490.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000490.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 207.5, "y1": 197.5, "x2": 286.0, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001198" }, { "image": "data/prediction/urban_night_rittenhouse/image/000490.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000490.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 207.5, "y1": 197.5, "x2": 286.0, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001198" }, { "image": "data/prediction/urban_night_rittenhouse/image/000491.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000491.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 151.0, "y1": 205.5, "x2": 268.0, "y2": 290.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001199" }, { "image": "data/prediction/urban_night_rittenhouse/image/000491.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000491.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 151.0, "y1": 205.5, "x2": 268.0, "y2": 290.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001199" }, { "image": "data/prediction/urban_night_rittenhouse/image/000492.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000492.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 48.5, "y1": 212.0, "x2": 245.0, "y2": 354.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001200" }, { "image": "data/prediction/urban_night_rittenhouse/image/000492.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000492.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 48.5, "y1": 212.0, "x2": 245.0, "y2": 354.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001200" }, { "image": "data/prediction/urban_night_rittenhouse/image/000493.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000493.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 294.5, "y1": 190.5, "x2": 346.0, "y2": 229.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001201" }, { "image": "data/prediction/urban_night_rittenhouse/image/000493.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000493.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 294.5, "y1": 190.5, "x2": 346.0, "y2": 229.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001201" }, { "image": "data/prediction/urban_night_rittenhouse/image/000494.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000494.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 309.5, "y1": 193.5, "x2": 361.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001202" }, { "image": "data/prediction/urban_night_rittenhouse/image/000494.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000494.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.5, "y1": 193.5, "x2": 361.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001202" }, { "image": "data/prediction/urban_night_rittenhouse/image/000495.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000495.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 299.0, "y1": 198.0, "x2": 361.5, "y2": 253.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001203" }, { "image": "data/prediction/urban_night_rittenhouse/image/000495.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000495.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.0, "y1": 198.0, "x2": 361.5, "y2": 253.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001203" }, { "image": "data/prediction/urban_night_rittenhouse/image/000496.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000496.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 292.5, "y1": 200.0, "x2": 373.0, "y2": 271.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001204" }, { "image": "data/prediction/urban_night_rittenhouse/image/000496.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000496.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 292.5, "y1": 200.0, "x2": 373.0, "y2": 271.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001204" }, { "image": "data/prediction/urban_night_rittenhouse/image/000497.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000497.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 378.5, "y1": 193.5, "x2": 460.5, "y2": 222.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001205" }, { "image": "data/prediction/urban_night_rittenhouse/image/000497.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000497.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 378.5, "y1": 193.5, "x2": 460.5, "y2": 222.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001205" }, { "image": "data/prediction/urban_night_rittenhouse/image/000498.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000498.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 340.0, "y1": 189.0, "x2": 414.5, "y2": 219.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001206" }, { "image": "data/prediction/urban_night_rittenhouse/image/000498.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000498.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 340.0, "y1": 189.0, "x2": 414.5, "y2": 219.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001206" }, { "image": "data/prediction/urban_night_rittenhouse/image/000499.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000499.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 484.0, "y1": 188.5, "x2": 577.0, "y2": 224.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001207" }, { "image": "data/prediction/urban_night_rittenhouse/image/000499.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000499.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 484.0, "y1": 188.5, "x2": 577.0, "y2": 224.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001207" }, { "image": "data/prediction/urban_night_rittenhouse/image/000500.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000500.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 215.0, "x2": 228.5, "y2": 387.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001208" }, { "image": "data/prediction/urban_night_rittenhouse/image/000500.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000500.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 215.0, "x2": 228.5, "y2": 387.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001208" }, { "image": "data/prediction/urban_night_rittenhouse/image/000501.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000501.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 221.5, "x2": 207.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001209" }, { "image": "data/prediction/urban_night_rittenhouse/image/000501.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000501.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 221.5, "x2": 207.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001209" }, { "image": "data/prediction/urban_night_rittenhouse/image/000502.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000502.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 223.5, "x2": 203.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001210" }, { "image": "data/prediction/urban_night_rittenhouse/image/000502.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000502.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 223.5, "x2": 203.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001210" }, { "image": "data/prediction/urban_night_rittenhouse/image/000503.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000503.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.0, "x2": 206.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001211" }, { "image": "data/prediction/urban_night_rittenhouse/image/000503.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000503.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.0, "x2": 206.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001211" }, { "image": "data/prediction/urban_night_rittenhouse/image/000504.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000504.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 221.5, "x2": 203.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001212" }, { "image": "data/prediction/urban_night_rittenhouse/image/000504.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000504.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 221.5, "x2": 203.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001212" }, { "image": "data/prediction/urban_night_rittenhouse/image/000505.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000505.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001213" }, { "image": "data/prediction/urban_night_rittenhouse/image/000505.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000505.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001213" }, { "image": "data/prediction/urban_night_rittenhouse/image/000506.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000506.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.0, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001214" }, { "image": "data/prediction/urban_night_rittenhouse/image/000506.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000506.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.0, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001214" }, { "image": "data/prediction/urban_night_rittenhouse/image/000507.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000507.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 222.0, "x2": 204.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001215" }, { "image": "data/prediction/urban_night_rittenhouse/image/000507.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000507.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 222.0, "x2": 204.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001215" }, { "image": "data/prediction/urban_night_rittenhouse/image/000508.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000508.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001216" }, { "image": "data/prediction/urban_night_rittenhouse/image/000508.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000508.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001216" }, { "image": "data/prediction/urban_night_rittenhouse/image/000509.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000509.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 222.0, "x2": 205.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001217" }, { "image": "data/prediction/urban_night_rittenhouse/image/000509.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000509.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 222.0, "x2": 205.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001217" }, { "image": "data/prediction/urban_night_rittenhouse/image/000510.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000510.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 418.0, "y1": 188.5, "x2": 525.0, "y2": 227.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001218" }, { "image": "data/prediction/urban_night_rittenhouse/image/000510.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000510.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 418.0, "y1": 188.5, "x2": 525.0, "y2": 227.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001218" }, { "image": "data/prediction/urban_night_rittenhouse/image/000511.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000511.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 354.0, "y1": 190.0, "x2": 466.5, "y2": 229.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001219" }, { "image": "data/prediction/urban_night_rittenhouse/image/000511.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000511.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 354.0, "y1": 190.0, "x2": 466.5, "y2": 229.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001219" }, { "image": "data/prediction/urban_night_rittenhouse/image/000512.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000512.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 315.0, "y1": 191.5, "x2": 409.0, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001220" }, { "image": "data/prediction/urban_night_rittenhouse/image/000512.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000512.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 315.0, "y1": 191.5, "x2": 409.0, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001220" }, { "image": "data/prediction/urban_night_rittenhouse/image/000513.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000513.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 312.5, "y1": 188.5, "x2": 368.0, "y2": 223.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001221" }, { "image": "data/prediction/urban_night_rittenhouse/image/000513.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000513.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 312.5, "y1": 188.5, "x2": 368.0, "y2": 223.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001221" }, { "image": "data/prediction/urban_night_rittenhouse/image/000514.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000514.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 264.0, "y1": 207.5, "x2": 393.5, "y2": 324.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001222" }, { "image": "data/prediction/urban_night_rittenhouse/image/000514.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000514.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 264.0, "y1": 207.5, "x2": 393.5, "y2": 324.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001222" }, { "image": "data/prediction/urban_night_rittenhouse/image/000515.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000515.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 245.5, "y1": 213.0, "x2": 409.5, "y2": 360.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001223" }, { "image": "data/prediction/urban_night_rittenhouse/image/000515.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000515.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 245.5, "y1": 213.0, "x2": 409.5, "y2": 360.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001223" }, { "image": "data/prediction/urban_night_rittenhouse/image/000516.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000516.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 244.0, "y1": 213.0, "x2": 415.5, "y2": 367.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001224" }, { "image": "data/prediction/urban_night_rittenhouse/image/000516.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000516.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 244.0, "y1": 213.0, "x2": 415.5, "y2": 367.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001224" }, { "image": "data/prediction/urban_night_rittenhouse/image/000517.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000517.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 245.5, "y1": 215.5, "x2": 416.0, "y2": 370.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001225" }, { "image": "data/prediction/urban_night_rittenhouse/image/000517.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000517.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 245.5, "y1": 215.5, "x2": 416.0, "y2": 370.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001225" }, { "image": "data/prediction/urban_night_rittenhouse/image/000518.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000518.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 243.0, "y1": 215.0, "x2": 416.0, "y2": 372.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001226" }, { "image": "data/prediction/urban_night_rittenhouse/image/000518.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000518.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.0, "y1": 215.0, "x2": 416.0, "y2": 372.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001226" }, { "image": "data/prediction/urban_night_rittenhouse/image/000519.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000519.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 246.0, "y1": 216.0, "x2": 418.5, "y2": 373.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001227" }, { "image": "data/prediction/urban_night_rittenhouse/image/000519.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000519.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 246.0, "y1": 216.0, "x2": 418.5, "y2": 373.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001227" }, { "image": "data/prediction/urban_night_rittenhouse/image/000520.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000520.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 241.0, "y1": 215.5, "x2": 414.0, "y2": 372.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001228" }, { "image": "data/prediction/urban_night_rittenhouse/image/000520.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000520.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 241.0, "y1": 215.5, "x2": 414.0, "y2": 372.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001228" }, { "image": "data/prediction/urban_night_rittenhouse/image/000521.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000521.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 241.0, "y1": 215.0, "x2": 413.0, "y2": 370.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001229" }, { "image": "data/prediction/urban_night_rittenhouse/image/000521.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000521.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 241.0, "y1": 215.0, "x2": 413.0, "y2": 370.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001229" }, { "image": "data/prediction/urban_night_rittenhouse/image/000522.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000522.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 243.5, "y1": 214.5, "x2": 414.5, "y2": 370.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001230" }, { "image": "data/prediction/urban_night_rittenhouse/image/000522.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000522.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.5, "y1": 214.5, "x2": 414.5, "y2": 370.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001230" }, { "image": "data/prediction/urban_night_rittenhouse/image/000523.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000523.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 243.0, "y1": 215.5, "x2": 414.5, "y2": 371.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001231" }, { "image": "data/prediction/urban_night_rittenhouse/image/000523.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000523.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 243.0, "y1": 215.5, "x2": 414.5, "y2": 371.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001231" }, { "image": "data/prediction/urban_night_rittenhouse/image/000524.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000524.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 559.0, "y1": 187.0, "x2": 639.5, "y2": 230.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001232" }, { "image": "data/prediction/urban_night_rittenhouse/image/000524.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000524.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 559.0, "y1": 187.0, "x2": 639.5, "y2": 230.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001232" }, { "image": "data/prediction/urban_night_rittenhouse/image/000525.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000525.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 500.0, "y1": 190.5, "x2": 621.5, "y2": 235.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001233" }, { "image": "data/prediction/urban_night_rittenhouse/image/000525.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000525.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 500.0, "y1": 190.5, "x2": 621.5, "y2": 235.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001233" }, { "image": "data/prediction/urban_night_rittenhouse/image/000526.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000526.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 434.0, "y1": 190.5, "x2": 559.0, "y2": 237.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001234" }, { "image": "data/prediction/urban_night_rittenhouse/image/000526.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000526.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 434.0, "y1": 190.5, "x2": 559.0, "y2": 237.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001234" }, { "image": "data/prediction/urban_night_rittenhouse/image/000527.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000527.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 380.0, "y1": 191.0, "x2": 506.0, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001235" }, { "image": "data/prediction/urban_night_rittenhouse/image/000527.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000527.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 380.0, "y1": 191.0, "x2": 506.0, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001235" }, { "image": "data/prediction/urban_night_rittenhouse/image/000528.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000528.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 338.0, "y1": 194.5, "x2": 456.0, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001236" }, { "image": "data/prediction/urban_night_rittenhouse/image/000528.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000528.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 338.0, "y1": 194.5, "x2": 456.0, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001236" }, { "image": "data/prediction/urban_night_rittenhouse/image/000529.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000529.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 299.5, "y1": 191.0, "x2": 403.0, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001237" }, { "image": "data/prediction/urban_night_rittenhouse/image/000529.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000529.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 299.5, "y1": 191.0, "x2": 403.0, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001237" }, { "image": "data/prediction/urban_night_rittenhouse/image/000530.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000530.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 296.0, "y1": 191.0, "x2": 356.0, "y2": 227.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001238" }, { "image": "data/prediction/urban_night_rittenhouse/image/000530.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000530.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 296.0, "y1": 191.0, "x2": 356.0, "y2": 227.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001238" }, { "image": "data/prediction/urban_night_rittenhouse/image/000531.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000531.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 527.5, "y1": 235.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001239" }, { "image": "data/prediction/urban_night_rittenhouse/image/000531.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000531.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 527.5, "y1": 235.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001239" }, { "image": "data/prediction/urban_night_rittenhouse/image/000532.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000532.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 469.5, "y1": 234.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001240" }, { "image": "data/prediction/urban_night_rittenhouse/image/000532.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000532.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 469.5, "y1": 234.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001240" }, { "image": "data/prediction/urban_night_rittenhouse/image/000533.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000533.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 232.5, "y1": 215.5, "x2": 404.0, "y2": 370.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001241" }, { "image": "data/prediction/urban_night_rittenhouse/image/000533.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000533.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 232.5, "y1": 215.5, "x2": 404.0, "y2": 370.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001241" }, { "image": "data/prediction/urban_night_rittenhouse/image/000534.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000534.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 240.0, "y1": 215.0, "x2": 411.0, "y2": 369.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001242" }, { "image": "data/prediction/urban_night_rittenhouse/image/000534.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000534.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 240.0, "y1": 215.0, "x2": 411.0, "y2": 369.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001242" }, { "image": "data/prediction/urban_night_rittenhouse/image/000535.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000535.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 241.5, "y1": 216.0, "x2": 413.0, "y2": 371.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001243" }, { "image": "data/prediction/urban_night_rittenhouse/image/000535.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000535.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 241.5, "y1": 216.0, "x2": 413.0, "y2": 371.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001243" }, { "image": "data/prediction/urban_night_rittenhouse/image/000536.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000536.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 238.5, "y1": 216.0, "x2": 409.5, "y2": 370.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001244" }, { "image": "data/prediction/urban_night_rittenhouse/image/000536.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000536.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 238.5, "y1": 216.0, "x2": 409.5, "y2": 370.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001244" }, { "image": "data/prediction/urban_night_rittenhouse/image/000537.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000537.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 242.5, "y1": 215.5, "x2": 413.5, "y2": 370.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001245" }, { "image": "data/prediction/urban_night_rittenhouse/image/000537.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000537.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 242.5, "y1": 215.5, "x2": 413.5, "y2": 370.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001245" }, { "image": "data/prediction/urban_night_rittenhouse/image/000538.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000538.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 241.5, "y1": 216.5, "x2": 412.0, "y2": 371.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001246" }, { "image": "data/prediction/urban_night_rittenhouse/image/000538.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000538.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 241.5, "y1": 216.5, "x2": 412.0, "y2": 371.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001246" }, { "image": "data/prediction/urban_night_rittenhouse/image/000539.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000539.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 242.5, "y1": 216.0, "x2": 414.5, "y2": 372.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001247" }, { "image": "data/prediction/urban_night_rittenhouse/image/000539.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000539.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 242.5, "y1": 216.0, "x2": 414.5, "y2": 372.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001247" }, { "image": "data/prediction/urban_night_rittenhouse/image/000540.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000540.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 242.0, "y1": 215.5, "x2": 413.5, "y2": 371.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001248" }, { "image": "data/prediction/urban_night_rittenhouse/image/000540.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000540.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 242.0, "y1": 215.5, "x2": 413.5, "y2": 371.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001248" }, { "image": "data/prediction/urban_night_rittenhouse/image/000541.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000541.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 238.0, "y1": 215.0, "x2": 409.5, "y2": 369.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001249" }, { "image": "data/prediction/urban_night_rittenhouse/image/000541.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000541.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 238.0, "y1": 215.0, "x2": 409.5, "y2": 369.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001249" }, { "image": "data/prediction/urban_night_rittenhouse/image/000542.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000542.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 237.0, "y1": 215.5, "x2": 408.5, "y2": 370.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001250" }, { "image": "data/prediction/urban_night_rittenhouse/image/000542.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000542.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 237.0, "y1": 215.5, "x2": 408.5, "y2": 370.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001250" }, { "image": "data/prediction/urban_night_rittenhouse/image/000543.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000543.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 451.0, "y1": 222.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001251" }, { "image": "data/prediction/urban_night_rittenhouse/image/000543.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000543.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 451.0, "y1": 222.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001251" }, { "image": "data/prediction/urban_night_rittenhouse/image/000544.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000544.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 440.5, "y1": 219.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001252" }, { "image": "data/prediction/urban_night_rittenhouse/image/000544.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000544.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 440.5, "y1": 219.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001252" }, { "image": "data/prediction/urban_night_rittenhouse/image/000545.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000545.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 439.5, "y1": 218.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001253" }, { "image": "data/prediction/urban_night_rittenhouse/image/000545.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000545.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 439.5, "y1": 218.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001253" }, { "image": "data/prediction/urban_night_rittenhouse/image/000546.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000546.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 440.5, "y1": 218.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001254" }, { "image": "data/prediction/urban_night_rittenhouse/image/000546.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000546.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 440.5, "y1": 218.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001254" }, { "image": "data/prediction/urban_night_rittenhouse/image/000547.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000547.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 439.0, "y1": 217.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001255" }, { "image": "data/prediction/urban_night_rittenhouse/image/000547.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000547.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 439.0, "y1": 217.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001255" }, { "image": "data/prediction/urban_night_rittenhouse/image/000548.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000548.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.5, "y1": 217.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001256" }, { "image": "data/prediction/urban_night_rittenhouse/image/000548.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000548.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.5, "y1": 217.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001256" }, { "image": "data/prediction/urban_night_rittenhouse/image/000549.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000549.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.5, "y1": 218.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001257" }, { "image": "data/prediction/urban_night_rittenhouse/image/000549.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000549.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.5, "y1": 218.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001257" }, { "image": "data/prediction/urban_night_rittenhouse/image/000550.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000550.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 437.5, "y1": 217.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001258" }, { "image": "data/prediction/urban_night_rittenhouse/image/000550.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000550.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 437.5, "y1": 217.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001258" }, { "image": "data/prediction/urban_night_rittenhouse/image/000551.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000551.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 438.0, "y1": 217.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001259" }, { "image": "data/prediction/urban_night_rittenhouse/image/000551.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000551.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 438.0, "y1": 217.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001259" }, { "image": "data/prediction/urban_night_rittenhouse/image/000552.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000552.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 214.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001260" }, { "image": "data/prediction/urban_night_rittenhouse/image/000552.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000552.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 214.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001260" }, { "image": "data/prediction/urban_night_rittenhouse/image/000553.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000553.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 226.0, "x2": 206.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001261" }, { "image": "data/prediction/urban_night_rittenhouse/image/000553.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000553.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 226.0, "x2": 206.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001261" }, { "image": "data/prediction/urban_night_rittenhouse/image/000554.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000554.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 222.0, "x2": 204.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001262" }, { "image": "data/prediction/urban_night_rittenhouse/image/000554.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000554.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 222.0, "x2": 204.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001262" }, { "image": "data/prediction/urban_night_rittenhouse/image/000555.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000555.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 226.5, "x2": 204.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001263" }, { "image": "data/prediction/urban_night_rittenhouse/image/000555.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000555.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 226.5, "x2": 204.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001263" }, { "image": "data/prediction/urban_night_rittenhouse/image/000556.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000556.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 223.0, "x2": 204.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001264" }, { "image": "data/prediction/urban_night_rittenhouse/image/000556.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000556.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 223.0, "x2": 204.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001264" }, { "image": "data/prediction/urban_night_rittenhouse/image/000557.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000557.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 223.0, "x2": 203.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001265" }, { "image": "data/prediction/urban_night_rittenhouse/image/000557.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000557.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 223.0, "x2": 203.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001265" }, { "image": "data/prediction/urban_night_rittenhouse/image/000558.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000558.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 225.0, "x2": 204.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001266" }, { "image": "data/prediction/urban_night_rittenhouse/image/000558.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000558.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 225.0, "x2": 204.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001266" }, { "image": "data/prediction/urban_night_rittenhouse/image/000559.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000559.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 226.0, "x2": 202.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001267" }, { "image": "data/prediction/urban_night_rittenhouse/image/000559.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000559.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 226.0, "x2": 202.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001267" }, { "image": "data/prediction/urban_night_rittenhouse/image/000560.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000560.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 227.5, "x2": 204.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001268" }, { "image": "data/prediction/urban_night_rittenhouse/image/000560.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000560.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 227.5, "x2": 204.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001268" }, { "image": "data/prediction/urban_night_rittenhouse/image/000561.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000561.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.0, "x2": 204.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001269" }, { "image": "data/prediction/urban_night_rittenhouse/image/000561.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000561.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.0, "x2": 204.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001269" }, { "image": "data/prediction/urban_night_rittenhouse/image/000562.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000562.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 206.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001270" }, { "image": "data/prediction/urban_night_rittenhouse/image/000562.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000562.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 206.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001270" }, { "image": "data/prediction/urban_night_rittenhouse/image/000563.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000563.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 565.0, "y1": 192.0, "x2": 639.5, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001271" }, { "image": "data/prediction/urban_night_rittenhouse/image/000563.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000563.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 565.0, "y1": 192.0, "x2": 639.5, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001271" }, { "image": "data/prediction/urban_night_rittenhouse/image/000564.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000564.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 460.0, "y1": 190.0, "x2": 575.0, "y2": 235.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001272" }, { "image": "data/prediction/urban_night_rittenhouse/image/000564.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000564.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 460.0, "y1": 190.0, "x2": 575.0, "y2": 235.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001272" }, { "image": "data/prediction/urban_night_rittenhouse/image/000565.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000565.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 375.5, "y1": 196.5, "x2": 486.5, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001273" }, { "image": "data/prediction/urban_night_rittenhouse/image/000565.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000565.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 375.5, "y1": 196.5, "x2": 486.5, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001273" }, { "image": "data/prediction/urban_night_rittenhouse/image/000566.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000566.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 342.5, "y1": 190.5, "x2": 412.5, "y2": 227.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001274" }, { "image": "data/prediction/urban_night_rittenhouse/image/000566.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000566.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 342.5, "y1": 190.5, "x2": 412.5, "y2": 227.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001274" }, { "image": "data/prediction/urban_night_rittenhouse/image/000567.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000567.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 428.5, "y1": 215.5, "x2": 639.5, "y2": 398.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001275" }, { "image": "data/prediction/urban_night_rittenhouse/image/000567.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000567.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.5, "y1": 215.5, "x2": 639.5, "y2": 398.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001275" }, { "image": "data/prediction/urban_night_rittenhouse/image/000568.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000568.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 421.0, "y1": 214.5, "x2": 636.0, "y2": 381.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001276" }, { "image": "data/prediction/urban_night_rittenhouse/image/000568.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000568.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.0, "y1": 214.5, "x2": 636.0, "y2": 381.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001276" }, { "image": "data/prediction/urban_night_rittenhouse/image/000569.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000569.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 415.0, "y1": 213.0, "x2": 620.5, "y2": 369.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001277" }, { "image": "data/prediction/urban_night_rittenhouse/image/000569.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000569.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 415.0, "y1": 213.0, "x2": 620.5, "y2": 369.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001277" }, { "image": "data/prediction/urban_night_rittenhouse/image/000570.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000570.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 410.0, "y1": 212.5, "x2": 610.5, "y2": 362.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001278" }, { "image": "data/prediction/urban_night_rittenhouse/image/000570.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000570.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 410.0, "y1": 212.5, "x2": 610.5, "y2": 362.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001278" }, { "image": "data/prediction/urban_night_rittenhouse/image/000571.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000571.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 408.5, "y1": 212.5, "x2": 605.5, "y2": 357.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001279" }, { "image": "data/prediction/urban_night_rittenhouse/image/000571.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000571.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.5, "y1": 212.5, "x2": 605.5, "y2": 357.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001279" }, { "image": "data/prediction/urban_night_rittenhouse/image/000572.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000572.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 404.5, "y1": 210.5, "x2": 587.0, "y2": 344.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001280" }, { "image": "data/prediction/urban_night_rittenhouse/image/000572.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000572.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 404.5, "y1": 210.5, "x2": 587.0, "y2": 344.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001280" }, { "image": "data/prediction/urban_night_rittenhouse/image/000573.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000573.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 385.0, "y1": 210.5, "x2": 525.5, "y2": 316.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001281" }, { "image": "data/prediction/urban_night_rittenhouse/image/000573.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000573.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 385.0, "y1": 210.5, "x2": 525.5, "y2": 316.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001281" }, { "image": "data/prediction/urban_night_rittenhouse/image/000574.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000574.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 375.5, "y1": 208.0, "x2": 480.5, "y2": 290.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001282" }, { "image": "data/prediction/urban_night_rittenhouse/image/000574.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000574.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 375.5, "y1": 208.0, "x2": 480.5, "y2": 290.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001282" }, { "image": "data/prediction/urban_night_rittenhouse/image/000575.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000575.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 359.0, "y1": 205.5, "x2": 448.0, "y2": 273.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001283" }, { "image": "data/prediction/urban_night_rittenhouse/image/000575.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000575.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 359.0, "y1": 205.5, "x2": 448.0, "y2": 273.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001283" }, { "image": "data/prediction/urban_night_rittenhouse/image/000576.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000576.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 373.0, "y1": 200.0, "x2": 443.5, "y2": 256.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001284" }, { "image": "data/prediction/urban_night_rittenhouse/image/000576.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000576.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 373.0, "y1": 200.0, "x2": 443.5, "y2": 256.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001284" }, { "image": "data/prediction/urban_night_rittenhouse/image/000577.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000577.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 240.0, "y1": 216.5, "x2": 409.0, "y2": 370.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001285" }, { "image": "data/prediction/urban_night_rittenhouse/image/000577.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000577.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 240.0, "y1": 216.5, "x2": 409.0, "y2": 370.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001285" }, { "image": "data/prediction/urban_night_rittenhouse/image/000578.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000578.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 238.5, "y1": 216.0, "x2": 407.0, "y2": 369.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001286" }, { "image": "data/prediction/urban_night_rittenhouse/image/000578.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000578.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 238.5, "y1": 216.0, "x2": 407.0, "y2": 369.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001286" }, { "image": "data/prediction/urban_night_rittenhouse/image/000579.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000579.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 240.5, "y1": 215.0, "x2": 410.0, "y2": 367.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001287" }, { "image": "data/prediction/urban_night_rittenhouse/image/000579.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000579.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 240.5, "y1": 215.0, "x2": 410.0, "y2": 367.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001287" }, { "image": "data/prediction/urban_night_rittenhouse/image/000580.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000580.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 242.0, "y1": 214.0, "x2": 405.5, "y2": 360.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001288" }, { "image": "data/prediction/urban_night_rittenhouse/image/000580.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000580.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 242.0, "y1": 214.0, "x2": 405.5, "y2": 360.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001288" }, { "image": "data/prediction/urban_night_rittenhouse/image/000581.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000581.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 246.0, "y1": 215.0, "x2": 401.5, "y2": 356.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001289" }, { "image": "data/prediction/urban_night_rittenhouse/image/000581.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000581.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 246.0, "y1": 215.0, "x2": 401.5, "y2": 356.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001289" }, { "image": "data/prediction/urban_night_rittenhouse/image/000582.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000582.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 249.5, "y1": 213.5, "x2": 393.0, "y2": 343.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001290" }, { "image": "data/prediction/urban_night_rittenhouse/image/000582.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000582.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 249.5, "y1": 213.5, "x2": 393.0, "y2": 343.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001290" }, { "image": "data/prediction/urban_night_rittenhouse/image/000583.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000583.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 261.0, "y1": 211.5, "x2": 383.0, "y2": 321.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001291" }, { "image": "data/prediction/urban_night_rittenhouse/image/000583.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000583.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 261.0, "y1": 211.5, "x2": 383.0, "y2": 321.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001291" }, { "image": "data/prediction/urban_night_rittenhouse/image/000584.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000584.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 273.5, "y1": 212.5, "x2": 374.5, "y2": 304.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001292" }, { "image": "data/prediction/urban_night_rittenhouse/image/000584.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000584.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 273.5, "y1": 212.5, "x2": 374.5, "y2": 304.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001292" }, { "image": "data/prediction/urban_night_rittenhouse/image/000585.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000585.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 268.0, "y1": 211.5, "x2": 363.0, "y2": 299.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001293" }, { "image": "data/prediction/urban_night_rittenhouse/image/000585.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000585.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 268.0, "y1": 211.5, "x2": 363.0, "y2": 299.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001293" }, { "image": "data/prediction/urban_night_rittenhouse/image/000586.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000586.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 288.5, "y1": 205.0, "x2": 385.0, "y2": 292.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001294" }, { "image": "data/prediction/urban_night_rittenhouse/image/000586.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000586.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 288.5, "y1": 205.0, "x2": 385.0, "y2": 292.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001294" }, { "image": "data/prediction/urban_night_rittenhouse/image/000587.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000587.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 225.5, "x2": 205.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001295" }, { "image": "data/prediction/urban_night_rittenhouse/image/000587.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000587.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 225.5, "x2": 205.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001295" }, { "image": "data/prediction/urban_night_rittenhouse/image/000588.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000588.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 223.5, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001296" }, { "image": "data/prediction/urban_night_rittenhouse/image/000588.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000588.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 223.5, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001296" }, { "image": "data/prediction/urban_night_rittenhouse/image/000589.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000589.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001297" }, { "image": "data/prediction/urban_night_rittenhouse/image/000589.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000589.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001297" }, { "image": "data/prediction/urban_night_rittenhouse/image/000590.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000590.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001298" }, { "image": "data/prediction/urban_night_rittenhouse/image/000590.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000590.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 205.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001298" }, { "image": "data/prediction/urban_night_rittenhouse/image/000591.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000591.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 223.5, "x2": 205.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001299" }, { "image": "data/prediction/urban_night_rittenhouse/image/000591.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000591.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 223.5, "x2": 205.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001299" }, { "image": "data/prediction/urban_night_rittenhouse/image/000592.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000592.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 206.0, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001300" }, { "image": "data/prediction/urban_night_rittenhouse/image/000592.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000592.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 224.5, "x2": 206.0, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001300" }, { "image": "data/prediction/urban_night_rittenhouse/image/000593.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000593.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 0.0, "y1": 229.5, "x2": 204.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001301" }, { "image": "data/prediction/urban_night_rittenhouse/image/000593.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000593.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 229.5, "x2": 204.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001301" }, { "image": "data/prediction/urban_night_rittenhouse/image/000594.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000594.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 232.0, "x2": 194.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001302" }, { "image": "data/prediction/urban_night_rittenhouse/image/000594.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000594.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 232.0, "x2": 194.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001302" }, { "image": "data/prediction/urban_night_rittenhouse/image/000595.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000595.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001303" }, { "image": "data/prediction/urban_night_rittenhouse/image/000595.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000595.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001303" }, { "image": "data/prediction/urban_night_rittenhouse/image/000596.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000596.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001304" }, { "image": "data/prediction/urban_night_rittenhouse/image/000596.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000596.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 0.0, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001304" }, { "image": "data/prediction/urban_night_rittenhouse/image/000597.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000597.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 441.5, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001305" }, { "image": "data/prediction/urban_night_rittenhouse/image/000597.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000597.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 441.5, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001305" }, { "image": "data/prediction/urban_night_rittenhouse/image/000598.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000598.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 428.5, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001306" }, { "image": "data/prediction/urban_night_rittenhouse/image/000598.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000598.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 428.5, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001306" }, { "image": "data/prediction/urban_night_rittenhouse/image/000599.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000599.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 433.0, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001307" }, { "image": "data/prediction/urban_night_rittenhouse/image/000599.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000599.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 433.0, "y1": 0.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001307" }, { "image": "data/prediction/urban_night_rittenhouse/image/000600.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000600.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 397.0, "y1": 201.0, "x2": 450.0, "y2": 248.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001308" }, { "image": "data/prediction/urban_night_rittenhouse/image/000600.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000600.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 397.0, "y1": 201.0, "x2": 450.0, "y2": 248.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001308" }, { "image": "data/prediction/urban_night_rittenhouse/image/000601.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000601.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 309.5, "y1": 208.0, "x2": 408.5, "y2": 297.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001309" }, { "image": "data/prediction/urban_night_rittenhouse/image/000601.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000601.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 309.5, "y1": 208.0, "x2": 408.5, "y2": 297.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001309" }, { "image": "data/prediction/urban_night_rittenhouse/image/000602.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000602.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 328.0, "y1": 207.0, "x2": 427.0, "y2": 294.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001310" }, { "image": "data/prediction/urban_night_rittenhouse/image/000602.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000602.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 328.0, "y1": 207.0, "x2": 427.0, "y2": 294.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001310" }, { "image": "data/prediction/urban_night_rittenhouse/image/000603.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000603.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 335.5, "y1": 207.5, "x2": 434.0, "y2": 295.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001311" }, { "image": "data/prediction/urban_night_rittenhouse/image/000603.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000603.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 335.5, "y1": 207.5, "x2": 434.0, "y2": 295.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001311" }, { "image": "data/prediction/urban_night_rittenhouse/image/000604.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000604.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 331.0, "y1": 207.0, "x2": 427.0, "y2": 292.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001312" }, { "image": "data/prediction/urban_night_rittenhouse/image/000604.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000604.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.0, "y1": 207.0, "x2": 427.0, "y2": 292.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001312" }, { "image": "data/prediction/urban_night_rittenhouse/image/000605.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000605.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 292.0, "y1": 202.5, "x2": 395.5, "y2": 286.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001313" }, { "image": "data/prediction/urban_night_rittenhouse/image/000605.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000605.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 292.0, "y1": 202.5, "x2": 395.5, "y2": 286.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001313" }, { "image": "data/prediction/urban_night_rittenhouse/image/000606.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000606.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 187.5, "y1": 204.5, "x2": 323.5, "y2": 292.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001314" }, { "image": "data/prediction/urban_night_rittenhouse/image/000606.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000606.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 187.5, "y1": 204.5, "x2": 323.5, "y2": 292.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001314" }, { "image": "data/prediction/urban_night_rittenhouse/image/000607.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000607.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 22.5, "y1": 199.5, "x2": 184.5, "y2": 296.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001315" }, { "image": "data/prediction/urban_night_rittenhouse/image/000607.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000607.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 22.5, "y1": 199.5, "x2": 184.5, "y2": 296.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001315" }, { "image": "data/prediction/urban_night_rittenhouse/image/000608.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000608.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 0.0, "y1": 211.0, "x2": 59.5, "y2": 310.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001316" }, { "image": "data/prediction/urban_night_rittenhouse/image/000608.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000608.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 0.0, "y1": 211.0, "x2": 59.5, "y2": 310.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001316" }, { "image": "data/prediction/urban_night_rittenhouse/image/000609.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000609.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 191.0, "y1": 196.5, "x2": 279.5, "y2": 269.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001317" }, { "image": "data/prediction/urban_night_rittenhouse/image/000609.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000609.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 191.0, "y1": 196.5, "x2": 279.5, "y2": 269.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001317" }, { "image": "data/prediction/urban_night_rittenhouse/image/000610.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000610.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 269.5, "y1": 208.5, "x2": 357.5, "y2": 288.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001318" }, { "image": "data/prediction/urban_night_rittenhouse/image/000610.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000610.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 269.5, "y1": 208.5, "x2": 357.5, "y2": 288.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001318" }, { "image": "data/prediction/urban_night_rittenhouse/image/000611.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000611.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 330.0, "y1": 204.0, "x2": 422.5, "y2": 288.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001319" }, { "image": "data/prediction/urban_night_rittenhouse/image/000611.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000611.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 330.0, "y1": 204.0, "x2": 422.5, "y2": 288.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001319" }, { "image": "data/prediction/urban_night_rittenhouse/image/000612.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000612.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 364.5, "y1": 207.5, "x2": 462.5, "y2": 296.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001320" }, { "image": "data/prediction/urban_night_rittenhouse/image/000612.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000612.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 364.5, "y1": 207.5, "x2": 462.5, "y2": 296.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001320" }, { "image": "data/prediction/urban_night_rittenhouse/image/000613.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000613.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 381.0, "y1": 207.0, "x2": 487.5, "y2": 303.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001321" }, { "image": "data/prediction/urban_night_rittenhouse/image/000613.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000613.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 381.0, "y1": 207.0, "x2": 487.5, "y2": 303.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001321" }, { "image": "data/prediction/urban_night_rittenhouse/image/000614.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000614.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 412.5, "y1": 211.0, "x2": 527.5, "y2": 317.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001322" }, { "image": "data/prediction/urban_night_rittenhouse/image/000614.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000614.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 412.5, "y1": 211.0, "x2": 527.5, "y2": 317.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001322" }, { "image": "data/prediction/urban_night_rittenhouse/image/000615.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000615.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 477.5, "y1": 220.5, "x2": 634.0, "y2": 352.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001323" }, { "image": "data/prediction/urban_night_rittenhouse/image/000615.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000615.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 477.5, "y1": 220.5, "x2": 634.0, "y2": 352.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001323" }, { "image": "data/prediction/urban_night_rittenhouse/image/000616.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000616.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 254.0, "y1": 187.5, "x2": 313.0, "y2": 228.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001324" }, { "image": "data/prediction/urban_night_rittenhouse/image/000616.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000616.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 254.0, "y1": 187.5, "x2": 313.0, "y2": 228.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001324" }, { "image": "data/prediction/urban_night_rittenhouse/image/000617.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000617.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 367.0, "y1": 203.5, "x2": 442.5, "y2": 259.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001325" }, { "image": "data/prediction/urban_night_rittenhouse/image/000617.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000617.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 367.0, "y1": 203.5, "x2": 442.5, "y2": 259.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001325" }, { "image": "data/prediction/urban_night_rittenhouse/image/000618.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000618.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 453.5, "y1": 203.0, "x2": 567.5, "y2": 285.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001326" }, { "image": "data/prediction/urban_night_rittenhouse/image/000618.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000618.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 453.5, "y1": 203.0, "x2": 567.5, "y2": 285.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001326" }, { "image": "data/prediction/urban_night_rittenhouse/image/000619.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000619.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 239.5, "y1": 183.0, "x2": 302.5, "y2": 217.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001327" }, { "image": "data/prediction/urban_night_rittenhouse/image/000619.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000619.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 239.5, "y1": 183.0, "x2": 302.5, "y2": 217.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001327" }, { "image": "data/prediction/urban_night_rittenhouse/image/000620.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000620.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 345.5, "y1": 198.5, "x2": 416.0, "y2": 243.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001328" }, { "image": "data/prediction/urban_night_rittenhouse/image/000620.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000620.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 345.5, "y1": 198.5, "x2": 416.0, "y2": 243.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001328" }, { "image": "data/prediction/urban_night_rittenhouse/image/000621.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000621.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 421.0, "y1": 201.5, "x2": 518.0, "y2": 265.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001329" }, { "image": "data/prediction/urban_night_rittenhouse/image/000621.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000621.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 421.0, "y1": 201.5, "x2": 518.0, "y2": 265.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001329" }, { "image": "data/prediction/urban_night_rittenhouse/image/000622.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000622.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 477.0, "y1": 207.5, "x2": 632.0, "y2": 307.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001330" }, { "image": "data/prediction/urban_night_rittenhouse/image/000622.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000622.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 477.0, "y1": 207.5, "x2": 632.0, "y2": 307.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001330" }, { "image": "data/prediction/urban_night_rittenhouse/image/000623.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000623.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 430.0, "y1": 196.5, "x2": 493.5, "y2": 242.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001331" }, { "image": "data/prediction/urban_night_rittenhouse/image/000623.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000623.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 430.0, "y1": 196.5, "x2": 493.5, "y2": 242.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001331" }, { "image": "data/prediction/urban_night_rittenhouse/image/000624.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000624.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 432.0, "y1": 201.0, "x2": 514.5, "y2": 255.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001332" }, { "image": "data/prediction/urban_night_rittenhouse/image/000624.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000624.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 432.0, "y1": 201.0, "x2": 514.5, "y2": 255.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001332" }, { "image": "data/prediction/urban_night_rittenhouse/image/000625.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000625.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 469.0, "y1": 212.0, "x2": 596.0, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001333" }, { "image": "data/prediction/urban_night_rittenhouse/image/000625.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000625.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 469.0, "y1": 212.0, "x2": 596.0, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001333" }, { "image": "data/prediction/urban_night_rittenhouse/image/000626.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000626.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 458.5, "y1": 200.5, "x2": 580.0, "y2": 279.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001334" }, { "image": "data/prediction/urban_night_rittenhouse/image/000626.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000626.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 458.5, "y1": 200.5, "x2": 580.0, "y2": 279.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001334" }, { "image": "data/prediction/urban_night_rittenhouse/image/000627.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000627.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 378.5, "y1": 191.5, "x2": 465.5, "y2": 222.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001335" }, { "image": "data/prediction/urban_night_rittenhouse/image/000627.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000627.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 378.5, "y1": 191.5, "x2": 465.5, "y2": 222.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001335" }, { "image": "data/prediction/urban_night_rittenhouse/image/000628.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000628.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 348.5, "y1": 192.5, "x2": 431.5, "y2": 226.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001336" }, { "image": "data/prediction/urban_night_rittenhouse/image/000628.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000628.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 348.5, "y1": 192.5, "x2": 431.5, "y2": 226.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001336" }, { "image": "data/prediction/urban_night_rittenhouse/image/000629.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000629.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 341.5, "y1": 190.5, "x2": 402.5, "y2": 227.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001337" }, { "image": "data/prediction/urban_night_rittenhouse/image/000629.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000629.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 341.5, "y1": 190.5, "x2": 402.5, "y2": 227.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001337" }, { "image": "data/prediction/urban_night_rittenhouse/image/000630.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000630.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 379.5, "y1": 169.5, "x2": 434.0, "y2": 217.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001338" }, { "image": "data/prediction/urban_night_rittenhouse/image/000630.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000630.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 379.5, "y1": 169.5, "x2": 434.0, "y2": 217.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001338" }, { "image": "data/prediction/urban_night_rittenhouse/image/000631.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000631.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 383.0, "y1": 167.5, "x2": 443.5, "y2": 221.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001339" }, { "image": "data/prediction/urban_night_rittenhouse/image/000631.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000631.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 383.0, "y1": 167.5, "x2": 443.5, "y2": 221.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001339" }, { "image": "data/prediction/urban_night_rittenhouse/image/000632.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000632.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 388.0, "y1": 167.5, "x2": 452.0, "y2": 224.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001340" }, { "image": "data/prediction/urban_night_rittenhouse/image/000632.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000632.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 388.0, "y1": 167.5, "x2": 452.0, "y2": 224.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001340" }, { "image": "data/prediction/urban_night_rittenhouse/image/000633.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000633.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 387.5, "y1": 167.0, "x2": 462.5, "y2": 228.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001341" }, { "image": "data/prediction/urban_night_rittenhouse/image/000633.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000633.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 387.5, "y1": 167.0, "x2": 462.5, "y2": 228.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001341" }, { "image": "data/prediction/urban_night_rittenhouse/image/000634.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000634.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 393.5, "y1": 173.0, "x2": 465.5, "y2": 232.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001342" }, { "image": "data/prediction/urban_night_rittenhouse/image/000634.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000634.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 393.5, "y1": 173.0, "x2": 465.5, "y2": 232.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001342" }, { "image": "data/prediction/urban_night_rittenhouse/image/000635.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000635.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 401.0, "y1": 173.0, "x2": 484.5, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001343" }, { "image": "data/prediction/urban_night_rittenhouse/image/000635.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000635.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 401.0, "y1": 173.0, "x2": 484.5, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001343" }, { "image": "data/prediction/urban_night_rittenhouse/image/000636.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000636.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 411.5, "y1": 169.0, "x2": 516.5, "y2": 251.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001344" }, { "image": "data/prediction/urban_night_rittenhouse/image/000636.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000636.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 411.5, "y1": 169.0, "x2": 516.5, "y2": 251.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001344" }, { "image": "data/prediction/urban_night_rittenhouse/image/000637.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000637.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 424.5, "y1": 168.0, "x2": 563.5, "y2": 274.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001345" }, { "image": "data/prediction/urban_night_rittenhouse/image/000637.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000637.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 424.5, "y1": 168.0, "x2": 563.5, "y2": 274.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001345" }, { "image": "data/prediction/urban_night_rittenhouse/image/000638.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000638.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 445.0, "y1": 156.0, "x2": 639.5, "y2": 307.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001346" }, { "image": "data/prediction/urban_night_rittenhouse/image/000638.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000638.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 445.0, "y1": 156.0, "x2": 639.5, "y2": 307.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001346" }, { "image": "data/prediction/urban_night_rittenhouse/image/000639.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000639.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 481.0, "y1": 131.5, "x2": 639.5, "y2": 371.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001347" }, { "image": "data/prediction/urban_night_rittenhouse/image/000639.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000639.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 481.0, "y1": 131.5, "x2": 639.5, "y2": 371.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001347" }, { "image": "data/prediction/urban_night_rittenhouse/image/000640.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000640.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 198.0, "x2": 498.5, "y2": 252.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001348" }, { "image": "data/prediction/urban_night_rittenhouse/image/000640.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000640.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 198.0, "x2": 498.5, "y2": 252.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001348" }, { "image": "data/prediction/urban_night_rittenhouse/image/000641.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000641.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 458.0, "y1": 200.5, "x2": 557.5, "y2": 272.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001349" }, { "image": "data/prediction/urban_night_rittenhouse/image/000641.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000641.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 458.0, "y1": 200.5, "x2": 557.5, "y2": 272.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001349" }, { "image": "data/prediction/urban_night_rittenhouse/image/000642.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000642.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 498.0, "y1": 205.0, "x2": 639.5, "y2": 318.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001350" }, { "image": "data/prediction/urban_night_rittenhouse/image/000642.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000642.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 498.0, "y1": 205.0, "x2": 639.5, "y2": 318.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001350" }, { "image": "data/prediction/urban_night_rittenhouse/image/000643.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000643.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 429.0, "y1": 197.5, "x2": 492.5, "y2": 243.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001351" }, { "image": "data/prediction/urban_night_rittenhouse/image/000643.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000643.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 429.0, "y1": 197.5, "x2": 492.5, "y2": 243.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001351" }, { "image": "data/prediction/urban_night_rittenhouse/image/000644.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000644.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 451.5, "y1": 198.0, "x2": 541.0, "y2": 255.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001352" }, { "image": "data/prediction/urban_night_rittenhouse/image/000644.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000644.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 451.5, "y1": 198.0, "x2": 541.0, "y2": 255.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001352" }, { "image": "data/prediction/urban_night_rittenhouse/image/000645.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000645.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 489.5, "y1": 201.0, "x2": 618.5, "y2": 280.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001353" }, { "image": "data/prediction/urban_night_rittenhouse/image/000645.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000645.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 489.5, "y1": 201.0, "x2": 618.5, "y2": 280.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001353" }, { "image": "data/prediction/urban_night_rittenhouse/image/000646.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000646.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 541.5, "y1": 211.0, "x2": 639.5, "y2": 339.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001354" }, { "image": "data/prediction/urban_night_rittenhouse/image/000646.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000646.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 541.5, "y1": 211.0, "x2": 639.5, "y2": 339.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001354" }, { "image": "data/prediction/urban_night_rittenhouse/image/000647.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000647.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 169.0, "y1": 189.0, "x2": 221.0, "y2": 229.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001355" }, { "image": "data/prediction/urban_night_rittenhouse/image/000647.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000647.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 169.0, "y1": 189.0, "x2": 221.0, "y2": 229.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001355" }, { "image": "data/prediction/urban_night_rittenhouse/image/000648.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000648.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 331.5, "y1": 191.0, "x2": 419.5, "y2": 239.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001356" }, { "image": "data/prediction/urban_night_rittenhouse/image/000648.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000648.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 331.5, "y1": 191.0, "x2": 419.5, "y2": 239.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001356" }, { "image": "data/prediction/urban_night_rittenhouse/image/000649.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000649.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 464.5, "y1": 190.5, "x2": 599.0, "y2": 254.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001357" }, { "image": "data/prediction/urban_night_rittenhouse/image/000649.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000649.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B RIGHT" } ], "bbox_2d": { "x1": 464.5, "y1": 190.5, "x2": 599.0, "y2": 254.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001357" }, { "image": "data/prediction/urban_night_rittenhouse/image/000650.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000650.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 414.5, "y1": 204.5, "x2": 473.5, "y2": 247.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001358" }, { "image": "data/prediction/urban_night_rittenhouse/image/000650.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000650.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.5, "y1": 204.5, "x2": 473.5, "y2": 247.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001358" }, { "image": "data/prediction/urban_night_rittenhouse/image/000651.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000651.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 423.5, "y1": 205.0, "x2": 504.5, "y2": 258.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001359" }, { "image": "data/prediction/urban_night_rittenhouse/image/000651.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000651.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 423.5, "y1": 205.0, "x2": 504.5, "y2": 258.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001359" }, { "image": "data/prediction/urban_night_rittenhouse/image/000652.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000652.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 406.0, "y1": 200.0, "x2": 463.0, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001360" }, { "image": "data/prediction/urban_night_rittenhouse/image/000652.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000652.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 406.0, "y1": 200.0, "x2": 463.0, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001360" }, { "image": "data/prediction/urban_night_rittenhouse/image/000653.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000653.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 436.0, "y1": 203.5, "x2": 514.0, "y2": 256.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001361" }, { "image": "data/prediction/urban_night_rittenhouse/image/000653.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000653.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 436.0, "y1": 203.5, "x2": 514.0, "y2": 256.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001361" }, { "image": "data/prediction/urban_night_rittenhouse/image/000654.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000654.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 486.0, "y1": 212.0, "x2": 621.0, "y2": 298.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001362" }, { "image": "data/prediction/urban_night_rittenhouse/image/000654.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000654.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 486.0, "y1": 212.0, "x2": 621.0, "y2": 298.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001362" }, { "image": "data/prediction/urban_night_rittenhouse/image/000655.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000655.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 410.0, "y1": 197.0, "x2": 460.0, "y2": 231.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001363" }, { "image": "data/prediction/urban_night_rittenhouse/image/000655.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000655.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 410.0, "y1": 197.0, "x2": 460.0, "y2": 231.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001363" }, { "image": "data/prediction/urban_night_rittenhouse/image/000656.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000656.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 434.5, "y1": 201.0, "x2": 504.5, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001364" }, { "image": "data/prediction/urban_night_rittenhouse/image/000656.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000656.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 434.5, "y1": 201.0, "x2": 504.5, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001364" }, { "image": "data/prediction/urban_night_rittenhouse/image/000657.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000657.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 479.0, "y1": 208.5, "x2": 591.5, "y2": 282.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001365" }, { "image": "data/prediction/urban_night_rittenhouse/image/000657.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000657.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 479.0, "y1": 208.5, "x2": 591.5, "y2": 282.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001365" }, { "image": "data/prediction/urban_night_rittenhouse/image/000658.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000658.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 495.0, "y1": 216.5, "x2": 639.5, "y2": 317.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001366" }, { "image": "data/prediction/urban_night_rittenhouse/image/000658.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000658.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 495.0, "y1": 216.5, "x2": 639.5, "y2": 317.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001366" }, { "image": "data/prediction/urban_night_rittenhouse/image/000659.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000659.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 408.0, "y1": 197.0, "x2": 467.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001367" }, { "image": "data/prediction/urban_night_rittenhouse/image/000659.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000659.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 408.0, "y1": 197.0, "x2": 467.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001367" }, { "image": "data/prediction/urban_night_rittenhouse/image/000660.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000660.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 440.0, "y1": 199.0, "x2": 524.0, "y2": 254.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001368" }, { "image": "data/prediction/urban_night_rittenhouse/image/000660.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000660.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 440.0, "y1": 199.0, "x2": 524.0, "y2": 254.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001368" }, { "image": "data/prediction/urban_night_rittenhouse/image/000661.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000661.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 498.0, "y1": 209.0, "x2": 639.5, "y2": 303.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001369" }, { "image": "data/prediction/urban_night_rittenhouse/image/000661.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000661.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 498.0, "y1": 209.0, "x2": 639.5, "y2": 303.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001369" }, { "image": "data/prediction/urban_night_rittenhouse/image/000662.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000662.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 425.0, "y1": 192.0, "x2": 491.5, "y2": 239.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001370" }, { "image": "data/prediction/urban_night_rittenhouse/image/000662.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000662.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 425.0, "y1": 192.0, "x2": 491.5, "y2": 239.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001370" }, { "image": "data/prediction/urban_night_rittenhouse/image/000663.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000663.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 459.5, "y1": 195.5, "x2": 560.0, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001371" }, { "image": "data/prediction/urban_night_rittenhouse/image/000663.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000663.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 459.5, "y1": 195.5, "x2": 560.0, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001371" }, { "image": "data/prediction/urban_night_rittenhouse/image/000664.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000664.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 507.0, "y1": 202.0, "x2": 639.5, "y2": 309.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001372" }, { "image": "data/prediction/urban_night_rittenhouse/image/000664.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000664.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 507.0, "y1": 202.0, "x2": 639.5, "y2": 309.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001372" }, { "image": "data/prediction/urban_night_rittenhouse/image/000665.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000665.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 356.5, "y1": 163.0, "x2": 417.0, "y2": 224.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001373" }, { "image": "data/prediction/urban_night_rittenhouse/image/000665.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000665.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 356.5, "y1": 163.0, "x2": 417.0, "y2": 224.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001373" }, { "image": "data/prediction/urban_night_rittenhouse/image/000666.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000666.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 368.5, "y1": 159.0, "x2": 448.5, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001374" }, { "image": "data/prediction/urban_night_rittenhouse/image/000666.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000666.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 368.5, "y1": 159.0, "x2": 448.5, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001374" }, { "image": "data/prediction/urban_night_rittenhouse/image/000667.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000667.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 370.5, "y1": 150.5, "x2": 489.5, "y2": 262.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001375" }, { "image": "data/prediction/urban_night_rittenhouse/image/000667.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000667.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 370.5, "y1": 150.5, "x2": 489.5, "y2": 262.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001375" }, { "image": "data/prediction/urban_night_rittenhouse/image/000668.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000668.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 387.0, "y1": 129.5, "x2": 562.5, "y2": 296.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001376" }, { "image": "data/prediction/urban_night_rittenhouse/image/000668.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000668.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 387.0, "y1": 129.5, "x2": 562.5, "y2": 296.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001376" }, { "image": "data/prediction/urban_night_rittenhouse/image/000669.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000669.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 401.5, "y1": 97.5, "x2": 639.5, "y2": 373.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001377" }, { "image": "data/prediction/urban_night_rittenhouse/image/000669.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000669.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 401.5, "y1": 97.5, "x2": 639.5, "y2": 373.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001377" }, { "image": "data/prediction/urban_night_rittenhouse/image/000670.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000670.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 369.0, "y1": 197.5, "x2": 423.5, "y2": 246.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001378" }, { "image": "data/prediction/urban_night_rittenhouse/image/000670.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000670.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 369.0, "y1": 197.5, "x2": 423.5, "y2": 246.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001378" }, { "image": "data/prediction/urban_night_rittenhouse/image/000671.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000671.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 375.5, "y1": 197.0, "x2": 448.0, "y2": 257.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001379" }, { "image": "data/prediction/urban_night_rittenhouse/image/000671.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000671.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 375.5, "y1": 197.0, "x2": 448.0, "y2": 257.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001379" }, { "image": "data/prediction/urban_night_rittenhouse/image/000672.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000672.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 395.0, "y1": 198.5, "x2": 477.5, "y2": 273.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001380" }, { "image": "data/prediction/urban_night_rittenhouse/image/000672.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000672.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 395.0, "y1": 198.5, "x2": 477.5, "y2": 273.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001380" }, { "image": "data/prediction/urban_night_rittenhouse/image/000673.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000673.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 414.0, "y1": 197.5, "x2": 528.0, "y2": 298.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001381" }, { "image": "data/prediction/urban_night_rittenhouse/image/000673.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000673.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 414.0, "y1": 197.5, "x2": 528.0, "y2": 298.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001381" }, { "image": "data/prediction/urban_night_rittenhouse/image/000674.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000674.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 435.0, "y1": 205.5, "x2": 590.5, "y2": 340.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001382" }, { "image": "data/prediction/urban_night_rittenhouse/image/000674.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000674.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 435.0, "y1": 205.5, "x2": 590.5, "y2": 340.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001382" }, { "image": "data/prediction/urban_night_rittenhouse/image/000675.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000675.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 451.5, "y1": 209.5, "x2": 639.5, "y2": 386.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001383" }, { "image": "data/prediction/urban_night_rittenhouse/image/000675.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000675.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 451.5, "y1": 209.5, "x2": 639.5, "y2": 386.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001383" }, { "image": "data/prediction/urban_night_rittenhouse/image/000676.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000676.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 465.5, "y1": 215.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001384" }, { "image": "data/prediction/urban_night_rittenhouse/image/000676.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000676.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 465.5, "y1": 215.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001384" }, { "image": "data/prediction/urban_night_rittenhouse/image/000677.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000677.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 470.0, "y1": 214.0, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001385" }, { "image": "data/prediction/urban_night_rittenhouse/image/000677.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000677.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 470.0, "y1": 214.0, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001385" }, { "image": "data/prediction/urban_night_rittenhouse/image/000678.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000678.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 313.5, "y1": 190.0, "x2": 365.5, "y2": 228.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001386" }, { "image": "data/prediction/urban_night_rittenhouse/image/000678.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000678.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 313.5, "y1": 190.0, "x2": 365.5, "y2": 228.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001386" }, { "image": "data/prediction/urban_night_rittenhouse/image/000679.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000679.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 313.0, "y1": 189.5, "x2": 372.0, "y2": 232.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001387" }, { "image": "data/prediction/urban_night_rittenhouse/image/000679.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000679.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 313.0, "y1": 189.5, "x2": 372.0, "y2": 232.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001387" }, { "image": "data/prediction/urban_night_rittenhouse/image/000680.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000680.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 298.5, "y1": 188.0, "x2": 374.0, "y2": 236.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001388" }, { "image": "data/prediction/urban_night_rittenhouse/image/000680.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000680.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 298.5, "y1": 188.0, "x2": 374.0, "y2": 236.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001388" }, { "image": "data/prediction/urban_night_rittenhouse/image/000681.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000681.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 262.5, "y1": 185.5, "x2": 360.0, "y2": 238.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001389" }, { "image": "data/prediction/urban_night_rittenhouse/image/000681.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000681.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 262.5, "y1": 185.5, "x2": 360.0, "y2": 238.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001389" }, { "image": "data/prediction/urban_night_rittenhouse/image/000682.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000682.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 194.5, "y1": 183.5, "x2": 312.0, "y2": 238.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001390" }, { "image": "data/prediction/urban_night_rittenhouse/image/000682.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000682.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 194.5, "y1": 183.5, "x2": 312.0, "y2": 238.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001390" }, { "image": "data/prediction/urban_night_rittenhouse/image/000683.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000683.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 455.0, "y1": 204.0, "x2": 553.0, "y2": 264.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001391" }, { "image": "data/prediction/urban_night_rittenhouse/image/000683.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000683.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 455.0, "y1": 204.0, "x2": 553.0, "y2": 264.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001391" }, { "image": "data/prediction/urban_night_rittenhouse/image/000684.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000684.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 477.0, "y1": 208.5, "x2": 598.0, "y2": 285.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001392" }, { "image": "data/prediction/urban_night_rittenhouse/image/000684.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000684.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 477.0, "y1": 208.5, "x2": 598.0, "y2": 285.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001392" }, { "image": "data/prediction/urban_night_rittenhouse/image/000685.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000685.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "D STOP" } ], "bbox_2d": { "x1": 511.0, "y1": 215.5, "x2": 639.5, "y2": 318.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001393" }, { "image": "data/prediction/urban_night_rittenhouse/image/000685.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000685.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 511.0, "y1": 215.5, "x2": 639.5, "y2": 318.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001393" }, { "image": "data/prediction/urban_night_rittenhouse/image/000686.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000686.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C DECELERATE" } ], "bbox_2d": { "x1": 151.5, "y1": 182.0, "x2": 216.0, "y2": 206.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001394" }, { "image": "data/prediction/urban_night_rittenhouse/image/000686.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000686.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 151.5, "y1": 182.0, "x2": 216.0, "y2": 206.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001394" }, { "image": "data/prediction/urban_night_rittenhouse/image/000687.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000687.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 113.0, "y1": 180.0, "x2": 183.0, "y2": 206.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001395" }, { "image": "data/prediction/urban_night_rittenhouse/image/000687.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000687.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 113.0, "y1": 180.0, "x2": 183.0, "y2": 206.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001395" }, { "image": "data/prediction/urban_night_rittenhouse/image/000688.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000688.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 394.5, "y1": 197.0, "x2": 479.5, "y2": 252.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001396" }, { "image": "data/prediction/urban_night_rittenhouse/image/000688.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000688.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 394.5, "y1": 197.0, "x2": 479.5, "y2": 252.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001396" }, { "image": "data/prediction/urban_night_rittenhouse/image/000689.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000689.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 398.0, "y1": 201.0, "x2": 503.0, "y2": 264.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001397" }, { "image": "data/prediction/urban_night_rittenhouse/image/000689.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000689.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 398.0, "y1": 201.0, "x2": 503.0, "y2": 264.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001397" }, { "image": "data/prediction/urban_night_rittenhouse/image/000690.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000690.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 393.5, "y1": 203.5, "x2": 509.5, "y2": 273.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001398" }, { "image": "data/prediction/urban_night_rittenhouse/image/000690.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000690.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 393.5, "y1": 203.5, "x2": 509.5, "y2": 273.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001398" }, { "image": "data/prediction/urban_night_rittenhouse/image/000691.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000691.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 354.5, "y1": 200.5, "x2": 491.0, "y2": 274.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001399" }, { "image": "data/prediction/urban_night_rittenhouse/image/000691.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000691.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 354.5, "y1": 200.5, "x2": 491.0, "y2": 274.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001399" }, { "image": "data/prediction/urban_night_rittenhouse/image/000692.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000692.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 196.5, "y1": 193.0, "x2": 358.0, "y2": 261.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001400" }, { "image": "data/prediction/urban_night_rittenhouse/image/000692.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000692.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 196.5, "y1": 193.0, "x2": 358.0, "y2": 261.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001400" }, { "image": "data/prediction/urban_night_rittenhouse/image/000693.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000693.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 64.5, "y1": 185.5, "x2": 223.0, "y2": 248.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001401" }, { "image": "data/prediction/urban_night_rittenhouse/image/000693.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000693.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 64.5, "y1": 185.5, "x2": 223.0, "y2": 248.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001401" }, { "image": "data/prediction/urban_night_rittenhouse/image/000694.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000694.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "B ACCELERATE" } ], "bbox_2d": { "x1": 454.0, "y1": 211.5, "x2": 639.5, "y2": 399.5 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001402" }, { "image": "data/prediction/urban_night_rittenhouse/image/000694.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000694.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 454.0, "y1": 211.5, "x2": 639.5, "y2": 399.5 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001402" }, { "image": "data/prediction/urban_night_rittenhouse/image/000695.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000695.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 382.5, "y1": 208.0, "x2": 639.5, "y2": 372.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001403" }, { "image": "data/prediction/urban_night_rittenhouse/image/000695.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000695.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "C STRAIGHT" } ], "bbox_2d": { "x1": 382.5, "y1": 208.0, "x2": 639.5, "y2": 372.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001403" }, { "image": "data/prediction/urban_night_rittenhouse/image/000696.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000696.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its speed behavior over the next 1 second from the ego-vehicle's perspective.\n\nSPEED: choose one from \"A\": \"KEEP\", \"B\": \"ACCELERATE\", \"C\": \"DECELERATE\", \"D\": \"STOP\", where:\n - KEEP means maintaining the current speed,\n - ACCELERATE means speeding up,\n - DECELERATE means slowing down,\n - STOP means the object is stationary.\n\nRespond in one line, e.g., 'B ACCELERATE'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A KEEP" } ], "bbox_2d": { "x1": 270.5, "y1": 197.0, "x2": 497.5, "y2": 318.0 }, "subtask": "speed", "original_id": "prediction/m3ed_test_prediction.json:001404" }, { "image": "data/prediction/urban_night_rittenhouse/image/000696.jpg", "event": "data/prediction/urban_night_rittenhouse/event/000696.npz", "conversations": [ { "from": "human", "value": "Suppose you are driving, and the following image is captured from the vehicle's front camera. The bounding box marks a target object in the scene.\nIn the given driving scene image, the object highlighted by the bounding box represents a detected dynamic agent. Based on its position and motion cues, predict its path behavior over the next 1 second from the ego-vehicle's perspective.\n\nPATH: choose one from \"A\": \"LEFT\", \"B\": \"RIGHT\", \"C\": \"STRAIGHT\", \"D\": \"UNKNOWN\", interpreted from the ego-vehicle's viewpoint:\n - LEFT means moving toward the left side of the ego vehicle,\n - RIGHT means moving toward the right side of the ego vehicle,\n - STRAIGHT means moving forward along the ego vehicle's direction,\n - UNKNOWN means the motion direction cannot be clearly determined.\n\nRespond in one line, e.g., 'C STRAIGHT'. Do not add punctuation or explanation." }, { "from": "gpt", "value": "A LEFT" } ], "bbox_2d": { "x1": 270.5, "y1": 197.0, "x2": 497.5, "y2": 318.0 }, "subtask": "path", "original_id": "prediction/m3ed_test_prediction.json:001404" } ]